This week I swept up behind myself.

I’m going to reflect on the impact of COVID-19 below because it feels important to do so. I’ll put it in a separate section so you can easily skip it.

We started a new sprint this week, and it came with a lot of carry-over (as is sometimes the case, nobody’s perfect). That’s led to picking our battles a bit more wisely, since we don’t want to increase the ticket queue. As such, I started the sprint with a bunch of technical upgrades and debt payoff.

A few more interesting tickets remain though. For one, we had a security incident over the weekend which I helped deal with on Monday. In the spirit of transparency you can read the details of our investigation online. It turned out to be (some level of) user error, but we’re putting additional safeguards in place to protect against similar issues in the future.

I also spent more time on our ever encompassing integration. AE and AS have done some amazing work at writing isolated, loosely coupled code and I helped look at some of the contract testing which is going exactly where we want it to.

On my own end, I updated the admin UI to show information about how the integration is performing. It was one of those satisfying bits of work that touches every part of our service in one small way. Hopefully next week we’ll have the bulk of the integration done and can just do tweaks to finish it off.

Isolation, quarantine and what that means for DigiDeps

I work in a distributed team, based across sites in Birmingham and Nottingham. We often work from home, some of us moreso than in the office. As such, the switch to home working has been fairly straightforward so far.

Tuesday was our first all-remote retrospective/planning and I think it went great. We’re still figuring out how to transfer our in-person practices to digital (particularly voting and anonymish sticky notes), but I reckon we got a good 90% of value out of that session.

As of next week though, we’ll have to contend with children. Most of the team have kids in the house now and will be acting primarily as carers. There’s no doubt this will affect our output/velocity/other business-y words.

And that’s fine, right? Most of the changes to our service aren’t more important than raising a child safely and securely. And any that are can easily get done by skeleton staff. The “nice to have”s can be pushed back without mercy, whilst we wait for this situation to stabilise a bit.

So my view (as not a delivery manager, not a senior dev, not a line manager, just an observer) is that we intentionally ramp down a little bit and manage our prioritisation more closely. We should focus on the things that need to happen first, then work on the more important bits, but be happy to let things slip.

I reckon that looks something like:

1. Keep the live service running
2. Be available to other teams who need support
3. Work on high priority tickets
4. Improve resilience/reduce cost
5. Work on low priority tickets

That’s a start. We need to identify external expectations and how things should be prioritised, but I think we need to be realistic, pragmatic and flexible in these challenging times.

Summary

• Investigated the security incident
• Updated admin UI to handle integration
• Helped write contract tests for integration point
• Upgraded GOV.UK Design System on all our products
• Tweaked npm audit to be more useful
  • We’ve had a spate of deep-chain vulnerabilities recently which just aren’t being fixed. We need to balance being aware of problems with being able to release.
• Various bits of pairing
• Set up automated deployment with GitLab CI and openconnect into a corporate firewall
  • I think I’ll have to write a blog post on this, it was really interesting and satisfying work