Fork me on GitHub
#z-do-not-post-here-old-ask-the-speaker
<
2019-06-26
>
Pascal Wallenius09:06:52

@tom.clark can you share more about how you use Sensu?

Pascal Wallenius09:06:52

@tom.clark can you share more about how you use Sensu?

Tom Clark09:06:04

Hey Pascal. Sure! Each of the product teams I mentioned operate their own "instance" of the platform, which is split into a "dev" side and a "prd" side. We run a sensu service in each of those halves, for each product

Tom Clark09:06:49

Which really helps with signal to noise ratio

Tom Clark09:06:13

Because a team only ever sees alerts for their product - and then only ever for the prd side or dev side

Tom Clark09:06:53

That 1:1 mapping continues to slack channels, email floats and pagerduty series

Tom Clark09:06:06

So again lots of convention over configuration

Tom Clark09:06:57

Means the estate is very predictable - if you know a product's name you can predict where all of its resources are

Tom Clark09:06:34

The Core team provide base config to monitor common bits

Tom Clark09:06:57

And the "field" platform engineers layer any product-specific bits over the top of that

Tom Clark09:06:24

Means that the field engineers can integrate upgrades from the core team without trampling their local changes

Tom Clark09:06:32

(sorry, this should've been in a thread)

Jeff Gallimore (CTIO - Excella)09:06:46

Good stuff @pascal.wallenius, right?

Marcos Trama09:06:31

@nithya_ruff good speak. Some insights to the metrics: 1) calculate how many calls to a open source item was made in a time frame 2) how many hours/effort would be spent to create that feature in house 3) divide one bu the other. Maybe this can be a way to achieve some kpi to track ;)

nithya ruff09:06:29

Marcos- good feedback. The time to build in-house or hours saved is a good one.

Pete Nuwayser10:06:04

Thank you @jayne for filling in for @genek on 54,000 tales

Macca11:06:24

@tom.clark great talk, very interesting to hear how you have evolved your approach and continue to do so... when you eval’d k8, why did you only review the AWS options? We have just been through a similar process at Coats and looked across all hyperscale providers... was it just to difficult to move from AWS to another due to the lack of containers...? Or as you hinted at in your talk minimal change to minimise risk...?

Macca11:06:24

@tom.clark great talk, very interesting to hear how you have evolved your approach and continue to do so... when you eval’d k8, why did you only review the AWS options? We have just been through a similar process at Coats and looked across all hyperscale providers... was it just to difficult to move from AWS to another due to the lack of containers...? Or as you hinted at in your talk minimal change to minimise risk...?

Matt McKee11:06:39

@sthorne would love to get your view on SRE approaches and if you believe that you can or should separate Platform and Application teams when it comes to SRE poc/implementations

Matt McKee11:06:39

@sthorne would love to get your view on SRE approaches and if you believe that you can or should separate Platform and Application teams when it comes to SRE poc/implementations

Ashish Agrawal13:06:35

@cs021v Maslach - You talked about breakdown of community as a challenge. How do we improve relationships at work? Especially when there are conflicts/disagreements. Most of the time leadership tries to fix the person than the situation.What is your advice to handle this in best interest of employees/organizations and move forward?

lottie14:06:48

@sthorne thanks for the talk. how do sres manage knowledge at a high level? can you tell us anything about your tools? prr brings it in to the sre team presumably, but do you put value on people knowing or ease of accessibility of knowledge? or both?

lottie14:06:48

@sthorne thanks for the talk. how do sres manage knowledge at a high level? can you tell us anything about your tools? prr brings it in to the sre team presumably, but do you put value on people knowing or ease of accessibility of knowledge? or both?

Ulf14:06:53

When talking about free space from work, what is your view of 4 days work weeks, Monday to Thursday 7,5 hours a day, vs a normal 5 day work week ? @nicolefv

Ulf14:06:53

When talking about free space from work, what is your view of 4 days work weeks, Monday to Thursday 7,5 hours a day, vs a normal 5 day work week ? @nicolefv

Ashish Agrawal16:06:16

@steve773, @nicolefv: How do we really build a culture of seek,solve and spread across a large organization? How do we bring Psychological safety into the employees to achieve this?

Ashish Agrawal16:06:16

@steve773, @nicolefv: How do we really build a culture of seek,solve and spread across a large organization? How do we bring Psychological safety into the employees to achieve this?

Steve Spear17:06:44

Thanks for the question of seeking and solving problems. Referring to the ADM Rickover case in my talk (built out in Chapter 5 of The High Velocity Edge) http://bit.ly/HVEChap1n4n5, here're the key points: 1: First, define the root cause of all problems as lack of understanding--ignorance. 2: Commit time to generating/soliciting feedback to find flaws in thinking (be they resident only in plans/designs or if they already evident in operating/doing) 3: When those flaws in thinking/doing are found, don't wave them away as "unimportant" or not a priority. Show how to swarm them to study them and solve them, demonstrating that doing so returns time and removes overburden from the people affected by the problem. We talk to this issue "committing time" to see and solving in chapter 6 and 7 of The High Velocity Edge and again with specific healthcare examples in

Ashish Agrawal17:06:55

@mik : How does the shift from Project to product work in outsourced s/w work? Since those will have definitive start and end of contract date and thus teams would form and eventually dismantle in a project fashion.

Ashish Agrawal17:06:55

@mik : How does the shift from Project to product work in outsourced s/w work? Since those will have definitive start and end of contract date and thus teams would form and eventually dismantle in a project fashion.

Steve Spear17:06:52

Thanks for listening to my preso and See To Solve description. Here's my end of preso ask: For the last five years, I've been working on a way to create a virtual Andon cord to make problems for people on the front lines more visible and then engage people so that they can more quickly solve those problems. Basically, I want the front line workers to get the help they need without having to go back to a computer, fill out a ticket with hundreds of fields, and then pray the cavalry arrives. We've had some great successes in manufacturing operations, in hospitals to help nurses and frontline care givers give better care to patients, and in pharmaceutical development to make sure they have the equipment they need available when they need it. So the help I’m looking for is this. If you know of any departments or groups in your organization where you think this might be useful I'd love to hear more about it. I’m looking for groups where you have 30-300 frontline workers who aren't in front of a computer all day, are typically mobile and moving around, and might even need help from other departments or groups outside their control. Just reach out to me on Slack or email and we can talk more. <mailto:Steve@HVELLC.com|Steve@HVELLC.com>