Closed steveoh closed 6 years ago
Thanks for contacting us, @steveoh. @rynowak FYI
It would also be useful to output the configuration the application is running as.
We know that metrics based checks are a thing people are going to want to write. Where a metric check is being healthy/unhealthy based on some threshold of data you can collect. I think our samples have an example for memory. But we don't have a plan to include too many of these by default yet. You guys giving us the same examples of checks you want would help us decide which ones should be in the box, based on actual usage and scenarios.
There is work like this happening: https://github.com/dotnet/designs/blob/master/accepted/cross-platform-performance-monitoring.md and I would also like to see what form that takes before committing to a bunch of checks. For example, if the event pipe type system described in the linked doc exists then a check that lets you examine things from the event pipe makes sense, and may cover a large set of otherwise specialized scenarios. I don't know that it will, but I want to see and in the mean time it shouldn't be too hard to copy some samples or have community members provide some checks. I am trying to be conservative about what we put in the base system initially. Does that make sense?
Thanks @glennc. I'm closing this issue as it seems this feedback is incorporated in an effort, which is still in planning phase.
I think it would be great to make general machine metrics easily available as a health check. I thought @DamianEdwards mentioned cpu/ram utilization etc on a community standup.
Example syntax
I really like the ability to add a connection string as a check. I have not tried the preview bits but is that available for things like npgsql, redis, etc?