Issue 071
This issue is sponsored by:
đź“– Guide to Data-Driven Engineering Teams
We analyzed 7 million commits from 85,000 engineers, and share how you can include data-driven insights to increase team velocity. A new wave of Engineering leaders are relying on deep git analytics and visibility to determine what’s working (what’s not), celebrate victories, and quickly identify bottlenecks and risks. Request a free trial.
Articles & News
My thoughts on this are kinda tangential to the article, but everyone loves a good rant, right? To quote Jeff Hodges, “A systems engineer without a good startup idea inevitably winds up doing monitoring.” Building time series databases is a hard problem so maybe this post will head some startups off at the pass should they decide, “I know! We’ll build yet another damn monitoring service!”
The RED Method: How to Instrument Your Services
I’ve bene teaching the USE and RED methods to clients lately and it’s been fascinating to see the impact it has on their monitoring. In my experience, the biggest challenge in monitoring isn’t alert fatigue, but knowing what needs instrumented and monitored to begin with. USE and RED make for a fantastic starting point.
O11ycast - Ep. #3, Distributed Systems with Paul Biggar of Dark and Ep. #4, Systems Administration with Adam Jacob of Chef
The latest episodes of O11ycast are out, this time talking with Paul Biggar, founder of CirleCI and Dark, and Adam Jacob, CTO of Chef.
The talk recordings from Datadog’s DASH conference in NYC in July are up. There’s lots of great talks in here.
SQL vs. Flux: Choosing the right query language for time-series data
I linked to Influx’s announcement of their new query language, Flux, a few weeks back. Seems that elicited a rebuttal from the folks at TimescaleDB on why SQL is the clear choice for a query language. Both sides make compelling arguments, but my bet is that Flux will see much more adoption by virtue of SQL being a bit of a pain in the ass for more complex queries (seriously, who writes sub-queries on a regular basis?)
Observations on Observability - #O11ycon report
I had the pleasure of attending the inaugural o11ycon in San Francisco, CA last week. Lemme say: it was pretty great. I’m not sure if I can really compare it to anything and do it justice. One of the more interesting bits is the kinds of companies that were present: companies with high data ingestion volume and customers/dependencies that are likely to do unexpected things to/with the service (eg, Nylas, who gave the open talk). Also, the tea was freaking amazing, thanks to Nuika Tea (try the Formula #1). Anyways, this trip report is by @anca, who did a great job capturing the event in text/photo.
The folks at Datadog have posted their latest exploration of metrics for a service, this time taking aim at the Microsoft IIS web server.
Events
Monitorama AMS 2018 - September 4-5, 2018 - Amsterdam, Netherlands
Are in Europe? Everyone’s favorite monitoring conference, Monitorama, is coming to a canal near you. Monitorama kindly sent over a discount code for Monitoring Weekly readers: use MLOVEWEEKLY at checkout for €100 off.
GrafanaCon LA - February 25-26, 2018 - Los Angeles, CA USA
The CFP is open too, until Oct 15th, 2018.
See you next week!
– Mike (@mike_julian) Monitoring Weekly Editor