If Devopsdays London was an office party, Patrick Dubois would be the Head of HR keeping an eye on proceedings, the guest speakers would be the live entertainment, and the Zero Turnaround guys would lacing the punch with more vodka and photocopying their body parts.
Zero Turnaround bought the fun, as evidenced by their stand which looked more like a booze section from an off licence, and this video here, which had me giggling in the back row like a naughty school kid:
Watch out Grammy Awards. Naturally, I felt I had to go and have a chat with these guys during one of the breaks and find out if I could blag any of their free booze…
“It’s all been taken, by competition winners” explained Simon from behind a massive heap of Guinness and Irish Whiskey. “We’re just, er, looking after this lot” said Neeme (Simon’s Zero Turnaround partner in crime) Yeah, sure.
No luck with the free booze, I thought maybe I could distract them with a question or two about devops, and then pinch some beer while they were deep in thought.
“So what does DevOps mean to you guys?” I asked, with half an eye on a bottle of Jamesons…
“A lot of it is about streamlining and bringing value to the customers sooner” Said Simon Maple, Tech Evangelist at ZT. And I suppose that’s actually what it’s ultimately about. Bridging the Dev and Ops gap is really about making us work as a team more smoothly, which leads to us being able to get stuff to our customers more quickly and more reliably. “But of course you need to right tools to do it – you have to automate across Dev and Ops” He added “This helps us remove bottlenecks”, which reminded me about my bottle of whiskey objective, so I asked another question by means of a diversion…
“Do you see a big crossover between DevOps and Agile” I asked. “Yes, DevOps is sort of like Agile breaking out of Dev and into Ops, but you can do devops stuff without being Agile” which again is true, but I imagine if you’re truly agile as a business, you’re almost certainly going to be doing “devops”.
Next, I decided to chuck in a question that caused a whole heap of opinionated discussion at devopsdays, namely “Is DevOps a job title?” (correct answer of course is – “what does it matter what your job title is? You can call yourself the Prince of Darkness for all I care. What matters is getting stuff done”), but Simon went for “I don’t think so, is Agile a job title?” which also made a lot of sense.
“So” I asked, “what do you see as the main challenges to DevOps?” I thought this questions was sufficiently fluffy to warrant an equally fluffy response, during which I could maybe accidentally knock one of the cans of Guinness into my bag…
“Well DevOps has well and truly gone viral now, so our main challenge is to make sure we do it right, or it’ll get a bad name” said Simon, as if he’d heard that particular question a million times before. No luck with the Guinness there then.
After about 5 minutes of chatting I was no closer to bagging myself any free beer than I was when I turned up, but I was having a good chat, and eventually we got around to talking about their products, JRebel and (the one I was more interested in) Live Rebel. I’d already seen their demo and it looked pretty cool to me. I’ll trial JRebel soon and put a review up here too, as I think people will be pretty impressed (I know I was), but that’s for another time. If you want to know more about their products, go see their website, duuh! For those too lazy to go to google and type in “Zero Turnaround”, here’s their website: http://zeroturnaround.com
For those even lazier than that, here’s a REALLY brief summary:
Live Rebel is a deployment orchestrator. It rolls up code and db changes into a release package and automates the delivery to whatever environment you like. It supports Java, Ruby, Python, PHP and Perl (officially) and Neeme also says it’ll work with .Net as well (not so officially). It also plugs nicely into most decent CI systems, such as Jenkins/Hudson and Bamboo (for which there are plugins) and has a one-stop management console for managing your servers and environments etc and so forth.
Anyway, I still had one final question for Simon and Neeme… “What has been your highlight of devopsdays then guys?” I asked. “Meeting the real creme de la creme of the industry and just being part of a great conference” came the reply. So, not meeting me then. Thanks guys.
Simon Maple is @sjmaple on twitter and Neeme is @nemecec. Follow them at your own peril.
By James Betteley
No time for that, I’m seriously late. I was leaving the office just as someone said “James, before you go…” and that was the end of any hopes I had of getting here on time.
It’s 6:40pm, I’ve finally got my sh1t together, and we’re off! In act that’s a lie, they were off ages ago. I’m so late there aren’t even any seats, so I’m sat in the corner at the back, like a proper Billy No-mates
6:41pm: Chris is talking about big balls of mud and how they’ve gone from that, to a much smaller ball (no mention of mud this time).
6:42: Slides are going by quicker than I can type! Chris is talking about the importance of moving away from a “blame culture”. I personally hate blame culture, I think it was the French who invented it. Arf! (sorry).
6:45pm: Ok, there’s a slide on how to get to Continuous Delivery, I’m going to pay attention to this one…
It says you need:
- Cross functional product focused teams
- A focus on technical debt
- Sit the team close to their clients
- actively remove blame culture
- focus on self improvement
- radiate metrics
- collect metrics on work in progress
After a quick coffee break it’s time to interrogate the suspects – It’s Q&A time!
7:01pm: How do you share “commonality of functionality”? Asks someone who likes words ending in “ality”. Service oriented architecture was apparently a big help responds someone from the 7digital posse (they are a posse by the way Chris has been joined by some 7digital reinforcements).
7.02pm: The next question is about metrics and how they collect them. Apparently they’re working on a logging system, but I’m guessing they also use CI and some live reporting tools which I probably missed in the earlier slides! Oops.
7.05pm: “What didn’t work?” Asks someone (my favorite question so far. I’m giving it a 7.5 out of ten). Trying to patch things up didn’t work. The dependency chain caused a pain. Acceptance Tests were a pain (haha!) Using UI stuff and a shared DB caused Acceptance Test issues, they say. I nod in agreement.
7:07pm: Someone says something about keeping environments the same being a challenge. They’re meant to be the same??? Where’s the fun in that?
7.10pm: A question on blame culture is next up, namely “How do you get rid of it?” By not telling on people! Also, having a dev manager who protects from above is handy.
It’s how you respond [to blame] that’s important, as that’s what sets the tone.
7:11pm: How did you make the culture change? Asks someone who wants to know how they made the culture change. It’s another good question, and one I’d really like to learn from. It’s all very well having a great culture, and there’s no denying its importance, but how do you make a culture change if it’s not ideal to start with? Sadly the answer isn’t straightforward. The posse reply with things like “Adopt Agile principles”, “tech manifesto” (which sounds cool), “self-organising teams”, “small steps” and “leading by example”. Followed up with “hire well”, “you need champions!” “do workshops”. Also, “the CTO is pretty cool”. Hmmmm, so no “click here to change your culture” button then?
7:16pm: The next question is a corker. It went a bit like: “Usually have to change architecture of system to support Continuous Delivery, but also sometimes the architecture of organisation as well. Did this happen?” That’s the winner so far. “No” comes the answer. Damn. At 7 digital there’s a focus on lack of hierarchy, so quite a flat structure. Not much change was needed then, obviously. I think the word “culture” came up as well, and not for the first time.
7:18pm: How have they managed to integrate Ops, asks the next person, clearly fresh from devopsdays. “We’re still learning” is the honest sounding response. Not that the all haven’t been honest sounding. They’ve started assigning Ops people to “the team”, by which I assume they mean the project team.
7:20pm: Someone wants to know if they had a shared goal between tech/ops and dev? I think the answer is yes. Basically Rob (who is the head of both ops and dev) became head of both ops and dev, which helped. He also created a tech manifesto and is toying with the idea of putting up some posters. When I was a kid I had a poster of Airwolf in my bedroom. Not sure if that’s going to help anyone though.
7:21pm: “Is there a QA on the team?” is the next question. Yarp (I’m paraphrasing). But the QA person is more of a coach – everyone is expected to do it, but they’re there to lead. No separate dev manger or QA manager – everyone’s one great big team (aaahhhh).
7:23pm: Somebody has asked how they handle support, and whether there’s a support team. I think the person who responds says there’s a “Systems team”, who get a text or call in the middle of the night. It seems a bit cruel that they wait until the middle of the night to text them but what do I know? Apparently the devs may also get involved, so that’s ok. There’s an on-call team, “but this is an area for improvement” they confess. Mainly it’s a case of “call someone!”, which I personally think is pretty good. But they do stress how there’s a focus on monitoring so that they can catch as many issues as possible before they become, er, issues, if you catch my drift. They said it much better than I can write it.
7:26pm: “How frequently did you deploy your big ball of mud compared to how frequently you do it now?” And that question goes to contestant number 2. It used to be one every 3 months, but they don’t measure how frequently they can deploy stuff any more because it’s that frequent.
(that’s just showing off). Improving the deploy mechanism was all-important. And changing the culture to shift to more frequent releases. That word “culture” again.
7:30pm: This question sounds like a plant: “How do you have time to test stuff if you deploy so often?” asks some cheeky 7digital employee hidden in the audience. I’m joking of course, it’s a nice question because it leads to a well executed answer: Chris basically explains that because they deploy so often, their releases are very small. Also, they’ve automated the hell out of everything.
7:31pm: Dave asks a really good question but I’m far too slow to keep up! It included the phrase “separation of concerns” so was probably too complicated for me to understand anyway.
7:40pm: There’s a question about schema changes. I reckon the answer will include the word “culture”. it does. Somehow.
If there was a word cloud for this Q&A session then “culture” would dwarf all the others. Something tells me that “culture shift” is important.
7:45pm: “How do you manage project accounting?” “We don’t” – No mention of culture!
7:46pm: Someone asks “If there’s a production issue, like an outage, who takes ownership?”. Nice one, who indeed does take ownership? “Everyone, we have a culture of shared ownership”. Gah! it’s all about culture!
7:47pm: “How do you decide what projects get green lighted?” asks some poor innocent from the back of the room (and no, it wasn’t me). Apparently this has nothing to do with Continuous Delivery (and all the other questions have?) and there’s nobody from the product team here so that question lands on stony ground.
7:52pm: Banos treads a fine line by asking a question dangerously close to the time when we’re meant to be heading to the pub, but just about gets away with it “what CI system do you use?” he asks, and the answer is (drum roll…..) Team City! Actually there was no drum roll, I made that up. Then interestingly they say that everyone is in charge of looking after Team city and that they just trust each other! Crazyness!
8.02pm: “I was told we finish at 8″ says Chris, and she’s bloody well right, there’s a pub nearby and some of us are thirsty.
So, in conclusion, 7digital know their Continuous Delivery from their TDD, and “culture” is the word of the evening. I’m off for beer with Banos, and the rest of the London Continuous Delivery gang!
Keep an eye out for #londoncd on twitter for news of the next London Continuous Delivery meetup, or go to the London CD website. Also follow the likes of @matthewpskelton, @AgileSteveSmith, @banoss and @davenolan for more Continuous Delivery goodness.
All good plans come in 3 phases:
Although I won’t be collecting any underpants, I’ll be following this basic template (with a couple of tweaks here and there) during an Agile adoption initiative I’m currently working on.
In the South Park episode (from which I have taken the picture above) the boys discover a bunch of underpant-stealing gnomes, who are collecting underpants as part of a grand plan to make profit. The gnomes claim to be business experts but none of them appears to know what phase 2 of the plan is. All they know is that their business model is based on collecting underpants, and so that’s what they’ll do.
Unfortunately, I have been witness to a couple of attempts at adopting agile which weren’t very dissimilar to the underpants gnomes’ business plan. Namely, a business starts “Adopting Agile”, usually driven by the development team, where they start doing stand-ups and using a sprint board (this is phase 1) and somehow they are surprised when this doesn’t suddenly start producing profit. Clearly, “becoming Agile” isn’t as simple as that.
Phase 1 – Collect business reasons (not underpants)
So you’re going Agile. Presumably you’ve determined that this is what you want, and what your customers need. If you haven’t done this yet then stop right there and ask yourself “Do I Need Go Agile?”. The answer might be “no”, but does needing to go Agile have to be the only reason? Maybe you just want to go agile to see what the fuss is all about, or to make your business more attractive to potential new employees.
So lets assume we’re going agile, and you have valid business reasons to do so. My first suggestion would be to make those business reasons highly visible. You have to outline the existing issues and how Agile can help to fix them. Mitchell and Webb once did a sketch about a toothbrush company who had to try to think of some gimmick to add to their toothbrushes in order to keep increasing their sales. They came up with the idea of “dirty tongue”. This is where microscopic “tonguanoids” build up, and basically result in social exclusion and a lack of sex. Their solution: to put bristles on the other side of the toothbrush so that people can brush their tongues while they brush their teeth. People will buy these toothbrushes despite the fact that “brushing your tongue makes you retch, everybody knows that”. The point I’m making, very badly, is that it’s a lot easier to sell things if people think that what they’re buying into will fix some very real, tangible issue.
The same goes with Agile. To get the buy-in you need to make your agile adoption a success, you’ll need to identify how “going Agile” is going to make life better for everyone concerned.
If the problem you’ve got is that you never ship software on time, or you constantly fail to deliver what the customer wants, then it’s fairly easy to “sell” agile as the solution. The concept of sprints are a doddle for everyone to understand, and they’ll love the idea that the customer will have regular interactions with the development team, and get to see regular progress in the demos. “Of course!” they’ll say “It’s so obvious, why didn’t I think of that before”. The business should easily be able to see how short, sharp sprints with an emphasis on “working software” will make it easier to deliver what the customer wants, and manage their expectations of when it’ll be ready.
But what if those aren’t the problems you need to solve?
What if your problem is quality? How do you convince the business that Agile will result in a higher quality product? It’s not quite so easy. Agile itself won’t deliver better quality, but the good practices you’ll have to implement in order to successfully be agile will help to improve your quality. I was thinking about this the other day because it’s exactly the problem I was faced with.
Agile isn’t going to make it easier to reliably test our software. But to be agile, we need to be able to build and deploy our project rapidly so that we can test it right there and then, not tomorrow, not next week, but right now, so that the testers and devs can work in tandem, building features and signing them off and moving on to the next one. We have to facilitate this in order to be agile, so as a byproduct of going agile we might have to invest in creating a new build and deployment system. And it has to be quick so it’ll have to be automated.
So we have an automated build and deployment system, but to be able to reliably test our features we’ll have to make sure the environments are reliable. We can throw people at this problem and dedicate a team to making sure our environments are clean and regularly audited, or we can automate all that as well! Fortunately there are numerous tools and good practices we can follow to do this, just take a look at Chef, Puppet, Vagrant, and VMWare as examples of tools for automating deployments of virtual machines, and the concepts of “infrastructure as code” for good practices. (of course, if your hardware isn’t already virtualised the first thing to do is see whether it can be, and if it really, honestly can’t, then look at tools like Norton Ghost and Powershell for ways of automating as much as you can).
“Agile” and “Improved Quality” might not be the most obvious bed partners, but the journey to becoming agile almost forces you to take steps which will naturally go towards improving your quality.
Hopefully you’ll have enough “sales material” to put forward a great case for agile – you can deliver exactly what the customer wants, to a higher quality, and you can manage their expectations in a way you could never do before. And that’s just scratching the surface of what Agile can do for a business, but for the purposes of keeping this post to a reasonable length, I’ll leave it at those 3 things!
Phase 2 – Pick the most appropriate project, and start doing Scrum
The sales pitch is over and now it’s time to start doing stuff. Make your life a lot easier by picking a project that has as many of the following features as possible:
- Smart developers and testers
- Isn’t suffering from a tonne of technical debt
- Has users who are happy to get involved in early & regular feedback
- Is small, new or yet to begin
If you’re taking on an existing project, a good idea at this point is to benchmark your existing processes. Consider trying to measure the following:
- How long does it take to get a single change from request through to production deployment?
- How much time and money does it cost to fix an issue on production?
- How many bugs do you typically find on your production code every month?
- How often do you deliver features that don’t satisfy the customer?
- How often do you deliver features after the deadline?
Measuring some of the things above is clearly non-trivial, but if you can find these stats somewhere, they’ll be very useful benchmarks for you in the future. When you can demonstrate that all of these metrics are improved in your new Agile process, god-like status will soon follow.
I recommend doing Scrum because it’s simple and has the most support in terms of people with experience, material (books, courses etc), and tools. It’s a good “framework” to get you started, and once you’ve had success, you can evolve into other methods, or incorporate them into Scrum (such as BDD, TDD etc).
At about this point you’ll need to do some
brainwashing training. The concept of doing analysis, design, development and testing all at the same time is going to sound absolutely bonkers to some people. Try your best to explain it to them, but don’t waste too much time on this – just crack on and make a start!
Most people will enjoy the experience of working in this “new” way, and the first few sprints will probably benefit from the fact that everyone is performing better simply because they feel more invigorated. Use this opportunity to promote scrum across the organisation.
In this phase, always maintain a focus on “the business” and not just on the technical team. It’s important that the business feels part of this new process or they’ll just see it as some crazy dev thing which doesn’t really affect them, and they won’t try to understand it. Business people might refer to this as “Promoting Synergy”, which I’ve just shoe-horned into this post so that I can add a picture from Lonely Island’s “Like A Boss” video. However, I do like to make a point of always highlighting the extra business value we’re delivering, and make sure the Product Managers (soon to be “Product Owners”) are involved all the way. They represent the traditional link between the customers and what we’re delivering, and so it’s essential that they understand the benefits of agile.
I was recently asked about the impact of “going agile” on a project’s release schedule, and when we would be able to deliver the features we’ve promised to the customers. It’s difficult to explain that we no longer know when we’ll deliver stuff, but at some point, people will have to realise that this is the wrong question. I prefer the idea of a rolling roadmap, which is continually reviewed and updated (as often as you can afford to do it, really). Rolling Roadmaps give the business, as well as the customers a good idea of our intentions, but it is very different to fixed dates on a release schedule, or a traditional yearly roadmap. Of course, everyone needs to understand that the main driver for our deliverables will be the customers, and what the customer wants will usually change over a shorter period than you expect. So for your new “Agile” project, try to work towards implementing a rolling roadmap culture, and move away from long-term fixed delivery dates (if you can).
One final note on Phase 2: Make it fun, and make it different.
Phase 3 – Improve
Agile promotes “fast feedback loops” all over the place: in development we get fast feedback on our code through Continuous Integration, with BDD we get fast feedback to the Product Owners/BAs and of course with our more frequent releases we get faster feedback from the customers. And so it is with our Agile processes as a whole. With short sprints and the clever use of retrospectives we can continually tinker with our fine tuning to see if we can improve our quality and velocity. Look at areas you can try to improve, change something and then see if your change has had a positive impact at the end of the sprint. This is basically the concept behind Deming’s Shewhart Cycle:
Deming actually preferred “Plan, Do, Study, Act”, whereas I myself prefer “Plan, Do, Measure, Act”. The reason I prefer this is because it implies the use of quantifiable metrics to base our actions on, rather than some other non-quantifiable observations. Anyways, the point is that after agile is applied, you should keep looking at ways to continuously improve. This is key to keeping everyone feeling fresh and invigorated, helps us to learn from our mistakes, and encourages innovation.
So there you go, Agile delivered in 3 well easy steps. It shouldn’t take you much longer than an afternoon. Ok, it might take a bit longer but if you’re looking for a 30,000 foot overview of a simple 3-phase approach, then you could do a lot worse than apply the principles of “Sell the Agile Idea, Pick the Best Project, and Keep Improving”.
The IT Ops team’s 4th sprint came to its conclusion recently… and we delivered all of our commitments! Huzzah! This post is basically a summary of how we analysed the sprint and what we got up to in the retrospective.
The main objectives of this sprint were:
- Rolling out a new password policy
- Rolling out a new file server
- Rebuilding the dev db
In addition to these tasks there were a bunch of other smaller tasks, far too dull for me to list here.
This was the first time we’d completed all of our committed points, and it gave us enough information to work out how many points we can realistically attempt to deliver in each sprint.
Once again the team completed over 100 points in the sprint, and once again we committed to doing 50 project points (the other 50-or-so being unplanned interruptions).
Sprint 4 in Numbers:
We completed 50 committed points
We completed 64 unplanned points
Total: 114 points
That’s very much in-line with the total we’ve come to expect.
We averaged 3.8 points per person per day for the duration of the sprint.
This means that we’re able to do roughly 2 points per person per day on committed project work.
In an ideal scenario, with no interruptions, no meetings and nothing to distract us from our tasks, we estimate that we could do about 6 points per person per day.
Therefore, statistically speaking, were actually 33% productive on our project work.
Of the remaining 67% we spent another 33% on interruptions.
So what happens to the other 33%??
The remaining effort and time is taken up with meetings, discussions, unrecorded interruptions, emails, phone calls and unproductive time. This figure (33%) is in the region of what I expected before we embarked upon this process some 8 weeks ago.
How to use this information
Well we know that we can complete approximately 50 project points per sprint. So we can now break down upcoming projects into tasks, and give realistic predictions on when they can be delivered.
The only caveat is that we do not make guarantees on project priority. We will always commit to working on delivering the highest business value projects first, and as a result will take guidance from the business on priorities. If a project’s priority is lowered then it will likely drop down the pecking order of our tasks, and its delivery date will slip.
In short, providing project priorities don’t chance for the duration of the project, we can now give far more realistic predictions on delivery dates.
To get the creative juices flowing, we started the retrospective with everyone drawing a picture of Sprint 4. There were no rules to this task, we could draw anything we wanted that represented sprint 4 for us as individuals.
Here’s the resulting “artwork”, if you can call it that:
Farid, who for some reason I expected to be something of an artist, drew a picture of the Jules Rimet trophy. This was partly to represent the fact that we completed the Password and File server projects, and partly to reflect that we hit our target.
Nick’s “effort” was of a question mark going into a grinder with a gold bar coming out of the other end. This reflects the fact that we’re not 100% sure of what’s happening up front at the beginning of our sprint, and we base our estimations on “best-guesses”, but by the end of the sprint we were producing the goods. It also raises a very relevant point about how we know we are addressing the right projects/tasks in the right order. More on that later.
Jon (in his unique avant-garde style) drew a person wearing a Boss hat. The person represents the team, and the boss hat is representative of the fact that we all completed our tasks on time and delivered what we committed to. Or at least that was his explanation, although I suspect it has more to do with the Lonely Island “Like a Boss” video, which is something of a theme tune in the IT Ops team. Seriously.
Leo drew the burndown, the original of which you can find at the bottom of this post.
Next up, we all chose 1 thing from the sprint that we wanted to discuss in more detail. Here’s what we ended up speaking about:
1. Using small post-it notes to record interruptions on the board didn’t work. They’re too small and not sticky enough. This was a major issue for Leo and Jon, who clearly have something against post-it notes.
2. We need to plan more carefully. Our planning sessions are rushed, and as a result we don’t think our tasks through as well as we should. The main failure here is that we don’t think about the impact other people can have on our tasks.
Following this, we discussed what we think went well, and what we think went badly in the sprint:
- Better prioritisation of interruptions allowed us to concentrate on the
- committed tasks
- Implementing the password policy
- Realistic goal setting
- IT request backlog grew bigger
What did we learn from Sprint 4:
- 50 Project points is our magic number
- Don’t over-commit
- Do more forward planning
- Our prioritizing is getting better
Finally we discussed the stats and figures mentioned earlier, and then we looked back at the burndown to see if there was any useful information we could take from it.
Leo suggested that the burndown be pre-populated with the mean line, so that we can see our daily deviation from that line. I have filed this suggestion in a folder named “Leo’s Good Ideas”, and we’ll adopt it in future sprints once I’ve worked out how to use Excel properly (or found a pen and a ruler).
As mentioned earlier, we’ve worked out what our average velocity is – and we can now confidently say that 50 project points per sprint is a realistic target for the team. Now that we’ve established this, we need to turn our attention to ensuring that we’re delivering the highest business value we possibly can. At present we feel we are slightly deficient in engaging with the business to define our priorities, so in the future we’ll be looking at having a much closer engagement with the rest of the business to help us with our prioritising and acceptance criteria. We want to be delivering the most important projects first, and we want to have a clear picture of what a successful delivery of those projects looks like.
Over the next couple of sprints we’ll be shifting our focus in this direction, and hopefully people will start to see the results over the next few weeks.
P.S. Jon is available for hire as a short order artist, ideal for weddings and birthday parties. Disappointment guaranteed.
As I write this, I’m sitting in a half-empty office in London. It’s half empty, you see, because it’s snowing outside, and when it snows in London, chaos ensues. Public transport grinds to a complete halt, buses just stop, and the drivers head for the nearest pub/cafe. The underground system, which you would think would be largely unaffected by snow, what with it being under ground, simply stops running. The overground train service has enough trouble running when it‘s sunny, let alone when it’s snowing. And of course most people know this, so whenever there’s a risk of snow, many people simply stay at home, hence the half-empty office I find myself in.
But why does London grind to such a standstill? Many northern European cities, as well as American ones, experience far worse conditions and yet life still runs fairly normally. Well, one reason for London’s regular winter shutdown is the infrastructure (you can see where I’m going with this, right?). The infrastructure in London is old and creaking, and in desperate need of some improvement. The problem is, it’s very hard to improve the existing infrastructure without causing a large amount of disruption, thus causing a great deal of inconvenience for the people who need to use it. The same can often be said about improving IT infrastructure.
Last night I went along to one of the excellent London Continuous Delivery Meetups (organised by Matthew Skelton at thetrainline.com – follow him on twitter here) which this month was all about Infrastructure Automation using Chef. Andy from Opscode gave us a demo of how to use Chef as part of a continuous delivery pipeline, which automatically provisioned an AWS vm to deploy to for testing. It all sounded fantastic, it’s exactly what many people are doing these days, it uses all the best tools, techniques and ideas from the world of continuous delivery, and of course, it didn’t work. There was a problem with the AWS web interface so we couldn’t actually see what was going on. In fact it looked like it wasn’t working at all. Anyway, aside from that slight misfortune, it was all very good indeed. The only problem is that it’s all a bit utopian. It would be great if we could all work on greenfield projects, or start rewriting everything from scratch, but in the real world, we often have legacy systems (and politics) which represent big blockers on the path to getting to utopia. I compare this to the situation with London’s Infrastructure – it’s about as “legacy” as you can possibly get, and the politics involved with upgrading it is obvious every time you pick up a newspaper.
In my line of work I’ve often come across the situation where new infrastructure was required – new build environments, new test server, new production environments and disaster recovery. In some cases this has been greenfield, but in most cases it came with the additional baggage of an existing legacy system. I generally propose one or more of the following:
- Build a new system alongside the old one, test it, and then swap it over.
- Take the old system out of commission for a period of time, upgrade it, and put it back online.
- Live with the old system, and just implement a new system for all projects going forward.
Then comes the politics. Sometimes there are reasons (budget, for instance) that prevents us from building out our own new system alongside the old one, so we’re forced into option 2 (by far the least favorable option because it causes the most amount of disruption).
The biggest challenge is almost always the Infrastructure Automation. Not from a technical perspective, but from a political point of view. It’s widely regarded as perfectly sensible to automate builds and deployments of applications, but for some reason, manually building, deploying and managing infrastructure is still widely tolerated! The first step away from this is to convince “management” that Infrastructure Automation is a necessity:
- Explain that if you don’t allow devs to log on to the live server to change the app code, then why is it acceptable to allow ops to go onto servers and change settings?
- Highlight the risk of human error when manually configuring servers
- Do some timings – how long does it take to manually build your infrastructure – from provisioning to handover (including any wait times for approval etc)? Compare this to how quick an automated system would be.
Once you’ve managed to convince your business that Infrastructure Automation is not just sensible, but a must-have, then it’s time for the easy part – actually doing it. As Andy was able to demonstrate (eventually), it’s all pretty straightforward.
Recently I’ve been using the cloud offerings from Amazon as a sort of stop-gap – moving the legacy systems to AWS, upgrading the original infrastructure by implementing continuous delivery and automating the infrastructure, and then moving the system back onto the upgraded (now fully automated and virtualised) system. This solution seems to fit a lot more comfortably with management who feel they’ve already spent enough of their budget on hardware and environments, and are loath to see the existing system go to waste (no matter how useless it is). By temporarily moving to AWS, upgrading the old kit and processes, and then swapping back, we’re ticking most people’s boxes and keeping everyone happy.
Cloud hosting solutions such as those offered by Amazon, Rackspace and Azure have certainly grown in popularity over the last few years, and in 2012 I saw more companies using AWS than I had ever seen before. What’s interesting for me is the way that people are using cloud hosting solutions: I am quite surprised to see so many companies totally outsourcing their test and production environments to the cloud, here’s why:
I’ve looked into the cost of creating “permanent” test labs in the cloud (with AWS and Rackspace) and the figures simply don’t add up for me. Building my own vm farm seems to make far more sense both practically and economically. Here are some figures:
3 Windows vms (2 webservers, 1 SQL server) minimum spec of dual core 4Gb RAM:
- 2x Windows “Large” instance
- 1x Windows “large” instance with SQL server
- Total: £432 ($693.20)
- 3x 4Gb dual core = £455
- 1x SQL Server = £o
- Total: £455
These figures assume a full 730 hours of service a month. With some very smart time and vm management you could get the rackspace cost down to about £300 pcm. However, their current process means you would have to actually delete your vms, rather than just power them off, in order to “stop the clock” so to speak.
So basically we’re looking at £450 a month for this simple setup. Of course it’s a lot cheaper if you go for the very low spec vms, but these were the specs I needed at the time, even for a test environment.
The truth is, for such a small environment, I probably could have cobbled together a virtualised environment of my own using spare kit in the server room, which would have cost next to nothing.
So lets look at a (very) slightly larger scale environment. The cost for an environment consisting of 8 Windows vms (with 1 SQL server) is around £1250 per month. After a year you would have spent £15k on cloud hosting!
But I can build my own vm farm with capacity for at least 50 vms for under £10k, so why would I choose to go with Rackspace or Amazon? Well, there are actually a few scenarios where AWS and Rackspace have come in useful:
1. When I just wanted a test environment up and running in no time at all – no need to deal with any ITOps team bottlenecks, just spin up a few vms and we’re away. In an ideal world, the infrastructure team should get a decent heads up when a new project is on it’s way, because the dev & QA team are going to need test environments setting up, and these things can sometimes take a while (more on that in a bit). But sadly, this isn’t an ideal world, and quite often the infrastructure team remain blissfully unaware of any hardware requirements until it’s blocking the whole project from moving forward. In this scenario, it has been convenient to spin up some vms on a hosted cloud and get the project unblocked, while we get on and build up the environments we should have been told about weeks ago (I’m not bitter, honestly )
2. Proof of concepting – Again no need to go through any red-tape, I can just get up and running on the cloud with minimal fuss.
3. When your test lab is down for maintenance/being rebuilt etc. If I could simply switch to a hosted cloud offering with minimal fuss, then I would have saved a LOT of downtime and emergencies in 2012. For example, at one company we hosted all our CI build servers on our own vm farm, and one day we lost the controller. We could have spun up another vm but for the fact that with one controller down, we were over capacity on the others. If I could have just spun up a copy of my Jenkins vm on AWS/Rackspace then I would have been back up and running in short order. Sadly, I didn’t have this option, and much panic ensued.
The Real Cost of Build-it-Yourself
So I’ve clearly been of the mind that hosting my own private cloud with a VMware VSphere setup is the most economically sensible solution. But is it really? What are the hidden costs?
Well last night, I was chatting with a couple of guys in the London Continuous Delivery community and they highlighted the following hidden costs of Build-it-Yourself (BIY):
Maintenance costs – With AWS they do the maintenance. Any hardware maintenance is done by them. In a BIY solution you have to spend the time and the money keeping the hardware ticking over.
Setup costs – Setting up a BIY solution can be costly. The upfront cost can be over £20,000 for a decent vm farm.
Management costs – The subsequent management costs can be very high for BIY systems. Who’s going to manage all those vms and all that hardware? You might (probably will) need to hire additional resources, that’s £40k gone!
So really, which solution is cheapest?
Our first ever ITOps sprint completed on Friday 7th December amid much fanfare and celebration (ok, that’s a bit of an exaggeration). The team completed a whopping grand total of 103 points. Among the highlights were:
- An Anti-Virus upgrade
- A new dev server
- 10+ IT Requests from the backlog
Yes, those were the highlights. But don’t scoff – that new dev server was AMAZEBALLS.
The format of our first retrospective was to spend 5 minutes writing down (on cards) what we felt went well in our first sprint, and what went not-so-well. We then discussed these and made a list of “lessons learned”. It turned out a bit like this:
|What went well||What went not-so-well|
|2 week “focus” made a huge difference: we could concentrate on a smaller number of tasks, rather than work from a massive backlog with no end in sight.||We under-estimated task sizes. Massively, in some cases.|
|Having the tasks on the board helped us see what everyone was doing||7 point task sizes were waaaaaaaaaaay too large|
|When working on a task, we no longer felt that we perhaps ought to be working on something else||We did fewer IT requests than we would otherwise have done|
|Time-boxing certain IT requests was a good idea.||We didn’t take holidays into account when we did our sprint planning! Duuuuuh.|
|Improved visibility: the rest of the world could see that we didn’t actually spend 2 weeks updating Facebook||We failed to get much done on the new File Server|
|We got through more project work than we expected|
|Easy to visualise the sheer amount of unplanned interruptions|
|We prioritised our interruptions quite well|
- 2 week sprints work well for the team’s focus
- We should time-box research into upcoming projects
- Any task which is above 5 points should be broken down into smaller tasks
- We need to improve our estimation
- We need to focus on completing the projects if we’ve committed to them
- We need to take holidays into account!
- Relative to interruptions, we can do more project work than we originally expected
Facts and Figures
- We completed 103 points
- 53 points of which were project points
- 50 points were interruptions
- 15 points were in progress at the end of the sprint
- 9 points (of committed work) remained incomplete
- 4.5 points were blocked
Friday to Tuesday were the least productive days in terms of project work. This was mainly due to
hangovers holidays which “we” (namely me) hadn’t taken into account, with 2/3rds of the team being off at one point.
Sprint 1 was an enjoyable success within the team. The increased visibility of interruptions as well as a clearer focus on our tasks was refreshing.
We failed to deliver one project (the new file server), which is disappointing, and we will need to focus on progressing and completing our project tasks in future.
We’ll continue with Scrum for the time being.
The other thing we noticed was that interruptions take up more time than just the duration of the interruption. There’s an additional amount of time on top of each interruption which needs to be taken into account. When someone is being productive working on a project task which, let’s say will take 3 hours, and they get interrupted to do an IT request which may only take 1 hour, the total time for completing the project task PLUS the interruption isn’t going to be 4 hours. It’s more likely to be 5, and this is because it takes some time to get back to being fully productive.
We made a conscious effort not to measure tasks by elapsed time, and instead we tried to use “Level of Effort”. I thought that mapping points to hours would be a bit too convenient for observers to draw incorrect conclusions by just adding up the number of hours and comparing it against the number of points we achieved! By using Level of Effort points and taking into account the number of interruptions (at the end of the sprint the points showed us that we’d done an almost equal number of interruption points as we had done project points), it was clear that in future we need to buffer our estimations somewhat.
It’s day 4 of our first agile ITOps sprint and so far it’s pretty much going to script. Our estimation of how much work we would be able to get through is looking roughly in the right ballpark (for more details on how we did the estimating, and other stuff see Agile ITOps: Day 1)
Looking at our sprint board a couple of things are fairly apparent. Firstly, that we were right about the number and frequency of interruptions, and secondly that we’re going to need a bigger board:
One of the things we aimed to achieve with the sprint board was to raise the visibility of the constant “interruptions” we have to work on. To do this we decided to note down every “interruption” on a red/pink card. As you can see in the picture above, we’ve worked more on interruptions than we have on our committed tasks. This was exactly what we expected to see.
Another thing we expected to see was a correlation between the number of interruptions we had, and the amount of project work we achieved. It was expected that on days where there were a high degree of interruptions, the number of project points completed would drop off. That hasn’t quite materialised yet, as you’ll see in the burndown below:
2 Plus 2 Doesn’t Equal 4?
Another thing that we’re expecting to see is that the maths just don’t add up. What I’m saying is that I don’t expect the team to be 100% productive when you add up the completed tasks we committed to deliver, plus all the points we did on interruptions. The reality is that people can’t automatically switch from one task to another and not lose any productivity. At the end of this sprint I’m hoping to be able to get a rough estimate of this “lost time”. In due course I will try to see if there’s any relationship between the amount of “lost time” and the types of tasks/interruptions we work on, so that we can effectively “cost” some of our regular tasks.
Today we started day 1 of our first ever ITOps sprint. This all came about because we needed a way of working out our productivity on “project tasks”, as well as learning how to triage our interruptions a bit better. None of the IT Ops team had tried any form of agile before, so there’s a mix of excitement and apprehension at the moment! I’ve opted for the old skool cards and sprint board (we’re using scrum by the way – more on this in a bit) rather than using an IT based system like Jira. This is mainly because everyone’s co-located, but also because I think it’s a good way to advertise what we’re doing within the office, and it’s an easy way to introduce a team to scrum.
The team already had a backlog of projects, so it was quite an easy job to break these down into cards. First of all we prioritised the projects and then sized up a bunch of tasks. We’re guessing, for this first sprint, that we’ll probably be around 40% to 50% productive on these project tasks, while the rest of our time will be taken up by interruptions (IT requests, responding to inquiries, dealing with unexpected issues etc).
I’m using a points system which is roughly equivalent to 1 point = 1 hour’s effort. This is a bit more granular than you might expect in a dev sprint, but since we’re dealing with a very high number of smaller tasks, I’ve opted for this slightly more granular scale.
As well as project work, there was also an existing list of IT requests which were in the queue waiting to be done. I wanted to “commit” to getting some of these done, so we planned them in to the sprint. I had originally thought of setting a goal of reducing the IT request queue by 30 tickets, and having that as a card, but I decided against this because I don’t feel it really represents a single task in it’s own right. I might change my mind about this at a later date though! So instead of doing that, we just chose the highest priority IT requests, and then arranged them by logical groupings (for instance, 3 IT requests to rebuild laptops would be grouped together). At the end of our planning session we had committed to achieving approximately 40% productivity.
We expect at least 50% of our time to be taken up by interruptions – this is based on a best-guess estimate, and will be refined from one sprint to the next as we progress on our agile journey. Whenever we get an interruption, let’s say someone comes over and says his laptop is broken, or we get an email saying a printer has mysteriously stopped working, then we scribble down a very brief summary of this issue on a red/pink card and make a quick call on whether it’s a higher priority that our current “in progress” tasks. If it is, then it goes straight into “in progress”, and we might have to move something out of “in progress” if the interruption means we need to park it for a while. Otherwise, the interruption will go in the backlog or in the “to do” list, depending on urgency.
At the end of the sprint we’re expecting to see a whole lot more red cards than green ones in the “Done” column. Indeed, we’re only a couple of hours into the first sprint an already the interruption cards contribute more than a 4:1 ratio in terms of hours actually worked today!
Triage and Priorities
One of the goals of this agile style of working is to get us to think a bit smarter when it comes to prioritising interruptions. It’s been felt in the past that interruptions always get given higher priority over our project work, when perhaps they shouldn’t have been. I suspect that in reality there’s a combination of reasons why the interruptions got done over project work: they’re usually smaller tasks, someone is usually telling you it’s urgent, it’s nice to help people out, and sometimes it’s not very easy to tell someone that you aren’t going to do their request because it’s not high enough priority. By writing the interruptions down on a card, and being able to look at the board and see a list of tasks we have committed to deliver, I’m hoping that we will be able to make better judgement calls on the priority, and also make it easier to show people that their requests are competing with a large number of other tasks which we’ve promised to get done.
I don’t actually see this as the perfect solution by any means, and as you might be able to tell already, there’s a slight leaning towards Kanban. I eventually want to go full-kanban, so to speak, but I feel that our scrum based approach provides a much easier introduction to agile concepts. Kanban seems to be much more geared towards an Ops style environment, with it’s continuous additions of tasks to the backlog and its rolling process. I think the pull based system and the focus on “Work In Progress” will provide some good tools for the ITOps team to manage their work flow. For now though, they’re getting down with a bit of scrum to get used to the terms and the ideas, and to understand what the hell the devs are talking about!
Retrospectives, burn-downs and Continuous Improvement
We’ll stick with scrum for a while and at the end of our sprint we’ll demo what project work we actually got through, as well as advertise the amount of IT requests we did, and the number of “interruptions” we successfully handled (although I don’t think I’ll keep calling them “interruptions” ). I’m also looking forward to our first retrospective, which we’ll be doing at the end of each sprint as a learning task – we want to see what we did well and what we did badly, as well as take a good look at our original estimations!
I’ll keep track of our burndown, and as the weeks go by I’ll try to give the team as much supporting metrics/information as I can, such as the number of hours since an interruption, the number of project tasks that are stuck in “In Progress” etc. But most important of all, we’re looking at this as a learning experience, an opportunity for us to improve the way we manage projects and a chance for everyone to get better at prioritising and estimating.
I’ll keep the posts coming whenever I get the chance!
Last night I was taken to the London premier of Warren Miller’s latest film called “Flow State”. Free beers, a free goodie bag and an hour and a half of the best snowboarders and skiers in the world doing tricks which in my head I can also do, but in reality are about 10000000% better than anything I can manage. Good times. Anyway, on my way home I was thinking about “flow” and how it applies to DevOps. It’s a tricky thing to maintain in an Ops capacity. It reminded me of a talk I went to last week where the speakers talked of the importance of “Flow” in their project, and it’s inspired me to write it up:
Thoughtworkers Pat and Aleksander have been working at a top secret location* for a top secret company** on a top secret mission to implement continuous delivery in a corporate Windows world***
* Ok, I actually forgot to ask where it was located
** It’s probably not a secret, but they weren’t telling me
*** It’s obviously not a secret mission seeing as how: a) it was the title of their talk and b) I just said what it was
Pat and Aleksander put their collective Powerpoint skills to good use and made a presentation on the stuff they’d done and learned during their time working on this top secret project, but rather than call their presentation “Stuff We Learned Doing a Project” (that’s what I would have named it) they decided to call it “Experience Report: Continuous Delivery in a Corporate Windows World”, which is probably why nobody ever asks me to come up with names for their presentations.
This talk was at Skills Matter in London, and on my way over there I compiled a list of questions which I was keen to hear their answers to. The questions were as follows:
- What tools did they use for infrastructure deployments? What VM technology did they use?
- How did they do db deployments?
- What development tools did they use? TFS?? And what were their good/bad points?
- Did they use a front-end tool to manage deployments (i.e. did they manage them via a C.I. system)?
- Was the company already bought-in to Continuous Delivery before they started?
- What breed of agile did they follow? Scrum, Kanban, TDD etc.
- What format were the built artifacts? Did they produce .msi installers?
- What C.I. system did they use (and why did they decide on that one)?
- Did they use a repository tool like Nexus/Artifactory?
- If they could do it all again, what would they do differently?
During the evening (mainly in the pub afterwards) Pat and Aleksander answered almost all of the questions above, but before I list them, I’ll give a brief summary of their talk. Disclaimer: I’m probably paraphrasing where I’m quoting them, and most of the content is actually my opinion, sorry about that. And apologies also for the completely unrelated snowboarding pictures.
Cultural Aspects of Continuous Delivery
Although CD is commonly associated with tools and processes, Pat and Aleksander were very keen to push the cultural aspects as well. I couldn’t agree more with this – for me, Continuous Delivery is more than just a development practice, it’s something which fundamentally changes the way we deliver software. We need to have an extremely efficient and reliable automated deployment system, a very high degree of automated testing, small consumable sized stories to work from, very reliable environment management, and a simplified release management process which doesn’t create problems than it solves. Getting these things right is essential to doing Continuous Delivery successfully. As you can imagine, implementing these things can be a significant departure from the traditional software delivery systems (which tend to rely heavily on manual deployments and testing, as well as having quite restrictive project and release management processes). This is where the cultural change comes into effect. Developers, testers, release managers, BAs, PMs and Ops engineers all need to embrace Continuous Delivery and significantly change the traditional software delivery model.
When Aleksander and Pat started out on this project, the dev teams were already using TFS as a build system and for source control. They eventually moved to using TeamCity as a Continuous Integration system, and Git-TFS as the devs primary interface to the source control system.
- Builds were done using msbuild
- They used TeamCity to store the build artifacts
- They opted for Nunit for unit testing
- Their build created zip files rather than .msi installers
- They chose Specflow for stories/spec/acceptance criteria etc
- Used Powershell to do deployments
- Sites were hosted on IIS
“Work in Progress” and “Flow” were the important metrics in this project by the sounds of things (suffice to say they used Kanban). I neglected to ask them if they actually measured their flow against quality. If I find out I’ll make a note here. Anyway, back to the project… because of the importance of Work in Progress and Flow they chose to use Kanban (as mentioned earlier) – but they still used iterations and weekly showcases. These were more for show than anything else: they continued to work off a backlog and any tasks that were unfinished at the end of one “iteration” just rolled over to the next.
Another point that Pat and Aleksander stressed was the importance of having good Business Analysts. They were essential in carving up stories into manageable chunks, avoiding “analysis paralysis”, shielding the devs from “fluctuating functionality” and ensuring stories never got stuck for too long. Some other random notes on their processes/practices:
- Used TDD with pairing
- Testers were embedded in the team
- Maintained a single branch of code
- Regression testing was automated
- They still had to raise a release request with Ops to get stuff deployed!
- The same artifact was deployed to every environment*
- The same deploy script was used on every environment*
* I mention these two points because they’re oh-so important principles of Continuous Delivery.
Obviously I approve of the whole TDD thing, testers embedded in the team, automated regression testing and so on, but not so impressed with the idea of having to raise a release request (manually) with Ops whenever they want to get stuff deployed, it’s not very “devops” I’d seek to automate that request/approval process. As for the “single branch of code”, well, it’s nice work if you can get it. I’m sure we’d all like to have a single branch to work from but in my experience it’s rarely possible. And please don’t say “feature toggling” at me.
One area the guys struggled with was performance testing. Firstly, it kept getting de-prioritised, so by the time they got round to doing it, it was a little late in the day, and I assume this meant that any design re-considerations they might have hoped to make could have been ruled out. Secondly, they had trouble actually setting up the load testing in Visual Studio – settings hidden all over the place etc.
Speaking with Pat, he was clearly very impressed with the wonders of Powershell scripting! He said they used it very extensively for installing components on top of the OS. I’ve just started using it myself (I’m working with Windows servers again) and I’m very glad it exists! However, Aleksander and Pat did reveal that the procedure for deploying a new test environment wasn’t fully automated, and they had to actually work off a checklist of things to do. Unfortunately, the reality was that every machine in every environment required some degree of manual intervention. I wish I had a bit more detail about this, I’d like to understand what the actual blockers were (before I run into them myself), and I hate to think that Windows can be a real blocker for environment automation.
Anyway, that’s enough of the detail, let’s get to the answers to the questions (I’ve added scores to the answers because I’m silly like that):
- What tools did they use for infrastructure deployments? What VM technology did they use? – Powershell! They didn’t provision the actual VMs themselves, the Ops team did that. They weren’t sure what tools they used. 1
- How did they do db deployments? – Pass 0
- What development tools did they use? TFS?? And what were their good/bad points? - TFS. Source Control and C.I. were bad so they moved to TeamCity and Git-TFS 2
- Did they use a C.I. tool to manage deployments? – Nope 0
- Was the company already bought-in to Continuous Delivery before they started? – They hired ThoughtWorks so I guess they must have been at least partly sold on the idea! Agile adoption was on their roadmap 1
- What breed of agile did they follow? Scrum, Kanban, TDD etc. – TDD with Kanban 2
- What format were the built artifacts? Did they produce .msi installers? – Negatory, they used zip files like any normal person would. 2
- What C.I. system did they use (and why did they decide on that one)? – TeamCity, which is interesting seeing as how ThoughtWorks Studios produce their own C.I. system called “Go”. I’ve used Go before and it’s pretty good conceptually, but it’s also expensive and hard to manage once you’re running over 50 builds and 100 test agents. The UI is buggy too. However, it has great features, but the Open Source competitors are catching up fast. 2
- Did they use a repository tool like Nexus/Artifactory? – They used TeamCity’s own internal repo, a bit like with Jenkins where you can store a build artifact. 1
- If they could do it all again, what would they do differently? – They wouldn’t push so hard for the Git TFS integration, it was probably not worth the considerable effort at the end of the day. 1
What does this total mean? Absolutely nothing at all.
What significance do all the snowboard pictures have in this article? None. Absolutely none whatsoever.
“As a topic of conversation, I find branching policies to be very interesting”, “Branching is great fun!”, “I wish we could do more branching” are just some of the comments on branching that you will never hear. And with good reason, because branching is boring. Merging is also boring. None of this stuff is fun. But for some strange reason, I still see the occasional branching policy which involves using the largest number of branches you can possibly justify, and of course the most random, highly complex merging process you can think of.
Here’s an example of a really $h!t branching policy:
Look how hateful it is!! I imagine this is the kind of conversation that leads to this sort of branching policy:
“Right, let’s just work on main and then take a release branch when we’re nearly ready to release”
“Waaaaait a second there… that sounds too easy. A better idea would be to have a branch for every environment, maybe one for each developer as well, and we should merge only at the most inconvenient time, and when we’ve merged to the production branch we should make a build and deploy it straight to Live, safe in the knowledge that the huge merge we just did went perfectly and couldn’t possibly have resulted in any integration issues”
“You see! Its complexity is beautiful”
Branching is boring. Merging is dull and risky. Don’t have more branches than you need. Work on main, take a branch at the latest possible time, release from there and merge daily. Don’t start conversations about branching with girls you’re trying to impress. Don’t talk about branches as if they have personalities, that’s just weird. Use a source control system that maintains branch history. Floss regularly. Stretch after exercise.