We’ve recently launched the OpenObservability Talks on June 25th, this is a continuation of the OpenObservability Conference we put together on May 27th. Following the ethos and discussions from the conference, we intend to talks, panels, and bits on Open-source Observability.
So if you’d like to watch the talk live on the last Thursday of the month, follow us at https://twitch.tv/OpenObservability or https://www.youtube.com/channel/UCLKOtaBdQAJVRJqhJDuOlPg and be part of the conversation.
If you miss the live session, we will be posting it at the YouTube channel as well as releasing it as a podcast, you can subscribe to it at https://anchor.fm/openobservability
Episode 1: Jack Neely
Our first episode had us hosting Jack Neely from 42 Lines where he gave his Five Tips on Application Monitoring with Prometheus talk. If you’d like to watch the recording it’s available here
Expanded from a lightning talk given at Monitorama PDX 2019, Jack gives five tips to open your mind about the power of Prometheus. It includes tricks to build dynamic threshold alerts, overcoming the limitations of the Prometheus histogram type for monitoring latency and events, and building SLO alerts that actually work.
A fun little quote and take away from the talk is this “Do you trust your service discovery? I don’t trust your service discovery? I don’t trust my service discovery! One of the fun problems in Prometheus is realizing services whenever failed there is no data, therefore, there’s new data to alert on pairing the up metric with the absent function is one of the few cases I advocate for.”
The next episode will be on July 30th, so keep an eye out for details on Twitter and if you have any ideas for what you’d like to see in future episodes please submit a talk to https://forms.gle/nQ5BHN8v6qcP9SaeA or email us at firstname.lastname@example.org about anything you’d like to see (or hear).