Aug 11, 2017
I came into this with the intention of being consistent with producing something weekly, but there has been a number of extenuating circumstances around actually doing this: between interviewing for promotion and planning a wedding, it's been rough so far as actually having the mental capacity to remember to do, well, anything goes.
Maybe I'm just scared of commitment.
Monday: Managing myself
I'm terrible at doing my timesheets: part of this is wilful disobedience purely because I've been told that I *must do them without ever been told of the benefit that me wasting half an hour allocating time to projects could possibly have; the other part is simply me forgetting about the process-oriented parts of my job.
I do do my timesheets often enough to have a favourite time code, though: the incomparable "managing myself." Monday was definitely a "managing myself" day, trying to get all of the paperwork I've been delaying around changing my name at work and changes to my pension that I haven't had time or been able to get done. Not exciting, but necessary.
One of my larger pieces of work at the moment is helping our compliance investigation division to define a flexible cloud environment that meets some fairly stringent legislative requirements around handling data, as well as this data potentially harmful materials, be they malware or in the realms of indecent or illegal content.
Only a small challenge, then. Certainly an interesting one.
Designing this in a greenfield environment, I certainly would agree with the approach that had been defined by the contractors employed to get this work going: their approach was dairly elegant, with isolation being provided by the use of dedicated VPCs in AWS; a key management service that wasn't reliant on KMS; and a development environment for data scientists that would prove incredibly flexible. We don't live in an ideal world, though: we have a model for the extension of our datacentres into the cloud that means that certain aspects of this design are, by necessity, delegated to another body in the department.
A day of whiteboarding later (I really should buy my own whiteboard markers - I've never been able to pick one up in a room and have it write well), a lot of explanations around why we designed the cloud networking the way we did; why we chose a vendor-agnostic strategy for any PaaS deployments; why complete control of the network environment couldn't be held by the project; why we mandate the technical controls that we do and we came to a compromise position that we believe will work, with work starting next week to spike this.
These sorts of results are what I really enjoy about my job: it's rare that contractors have any respect for permanent civil servants when it comes to anything technical (and our management doesn't help that - but that's really a subject to be discussed over a few drinks and not somewhere I could held to account for what I say), but having such a constructive conversation is wonderful.
Wednesday: "How do you know what's up or down?"
Partially on leave, partially trying to sell a customer on a monitoring solution for a large and complex distributed service, Wednesday was a bit of a wash.
The monitoring solution I'm trying to define for this service is premised on a few core tenets:
- Don't give a large vendor any money, please
- Take a modular approach so we can more easily move with the times as certain components become outmoded
- Don't use Oracle Enterprise Manager if you can help it
- Rely on guages, counters and log data rather than runtime instrumentation to get the required data
I concede fairly frequently that I inhabit a position in my job that is somewhat divorced from the reality of delivery: I sit in my ivory tower and declare what, from a corporate point of view, is and isn't acceptable. I welcome input from anyone willing to talk to me, but I have decision-making responsibilities wider than any one project or programme, which means that I don't necessarily have to worry about the first-adopter penalties associated with potentially corporate-wide services or cost of initial development. The programme are fairly resistant to any approach that isn't 'throw money at the problem, make it go away', but I think we've made some progress insomuch as they've at least agreed to do a technical spike for the stack (Sensu, InfluxDB, ELK for anyone interested).
This programme, however, is very much the opposite of my Tuesday conversation as far as the contractors go: there's a far more bullish, borderline xenophobic attitude to anything that is considered 'outside' of programme delivery. That certainly is a pattern of behaviour that concerns me: I'm all for camaraderie, but that sort of negative cohesion always worries me.
Thursday: An overdue team meeting
I'm lucky to work within the calibre of team that I work in: everyone is supportive of everyone else and we're generally left to do our own work and follow our own passions where time allows. We're in the process or revising how we make our work product (solution designs, technology catalogues, deployment patterns, best practice) available outside of our team and more widely across the department, and the technology catalogue application I've developed seems to have gone down well across our trial users and has cut down on a lot of low-value communication. I feel like we're at the vanguard with this sort of stuff as far as the the wider group's work goes, so that's certainly a good feeling.
Friday: Bringing MIS to the 21st Century
A fairly slow day, but I think I may have designed something to replace an ageing operational monitoring platform with something a lot cheaper and a lot more responsive. Hopefully, this will replace my party piece in interviews of 'I delivered a project ahead of time and below cost in a government department once, who else has done that?' Time will tell.
Music: Russian Red, Chelsea Wolfe, Sarah Fimm
Games: DOTA2, Guilty Gear Xrd Revelator, Battle Bakraid
Thought for the week
Repeating myself gets results eventually: it's recognising cultural differences between Telford and Southend that will speed that up in future. We're a weird organisation.