Setting up a custom Honeycomb dashboard for the opencodelists environment could allow us to better monitor and investigate things that matter to us. Suggest doing some basic initial setup that can be extended in future rather than trying to be comprehensive. We will learn more about our needs over time.
For example:
Overall traffic, average request delay, resource usage. This allows us to monitor overall load and performance.
Breakdown by http.target / http.route. This allows us to see which views are being accessed more easily, perhaps mapping to the underlying views, similar to the derived columns approach in #2131.
Specific error types we care most about, such as 404 or 500 errors.
Bot. vs non-bot traces.
Domain-specific such as breakdown by codelist interacted with, per codelist author, per user...
Setting up a custom Honeycomb dashboard for the
opencodelistsenvironment could allow us to better monitor and investigate things that matter to us. Suggest doing some basic initial setup that can be extended in future rather than trying to be comprehensive. We will learn more about our needs over time.For example:
http.target/http.route. This allows us to see which views are being accessed more easily, perhaps mapping to the underlying views, similar to the derived columns approach in #2131.For inspiration see these dashboards for Job Server.
Here is the default dashboard on opencodelists.