chore: IIS Modernization #1542
chore: IIS Modernization #1542
Conversation
| "uid": "${prometheus_datasource}" | ||
| }, | ||
| "expr": "rate(windows_iis_files_sent_total{job=~\"$job\", instance=~\"$instance\", site=~\"$site\"}[$__rate_interval])", | ||
| "expr": "increase(windows_iis_files_sent_total{job=\"integrations/iis\",job=~\"$job\",instance=~\"$instance\", site=~\"$site\"}[$__interval:] offset -$__interval)", |
There was a problem hiding this comment.
Changed this to increase as a file/seconds seems a tad less useful that just knowing the files transferred
Dasomeone
left a comment
Dasomeone
left a comment
There was a problem hiding this comment.
Only comment so far here is to perhaps reuse the common-lib panels a tad more, but overall I'm happy enough with this as-is.
It's cookie-cutter yes, and I remember that was something we discussed originally as well (same as some panels being "noisy" but we actually want to show 0 values in this one)
I have no real suggestions for improvements here, aside from potentially duplicating some panels as stat panels for a top row, but let's discuss in our sync
microsoft-iis-mixin/panels.libsonnet
Outdated
|
|
||
| // Overview - Requests | ||
| requests: | ||
| commonlib.panels.generic.timeSeries.base.new( |
There was a problem hiding this comment.
Any chance this could be converted to the common requests panel? https://github.com/grafana/jsonnet-libs/tree/master/common-lib/common/panels/requests/timeSeries
microsoft-iis-mixin/panels.libsonnet
Outdated
| + g.panel.timeSeries.options.legend.withAsTable(true), | ||
|
|
||
| requestErrors: | ||
| commonlib.panels.generic.timeSeries.base.new( |
There was a problem hiding this comment.
Same here,there's a preset for request errors https://github.com/grafana/jsonnet-libs/tree/master/common-lib/common/panels/requests/timeSeries
aalhour
left a comment
aalhour
left a comment
There was a problem hiding this comment.
I have a few requests for changes 🙏
- There is a bug with an existing alert that you didn't touch and I added it to the issue https://github.com/grafana/cloud-onboarding/issues/9972 as an example.
- There are a few instances of
increase(...)function not using the proper interval with offset, i.e.:[$__interval:] offset $__interval - There are a few counter signals missing the
rangeFunctionbut a couple of others have it - There is a potential issue in the
config.libsonnetfile - A few other minor comments
| blockedAsyncIORequests: { | ||
| name: 'Blocked async I/O requests', | ||
| type: 'counter', | ||
| rangeFunction: 'increase', |
| // Requests signals | ||
| requests: { | ||
| name: 'Requests per second', | ||
| type: 'counter', |
There was a problem hiding this comment.
Does this counter signal need a rangeFunction as well, similar to blockedAsyncIORequests?
|
|
||
| lockedErrors: { | ||
| name: 'Locked errors per second', | ||
| type: 'counter', |
There was a problem hiding this comment.
Does this counter signal need a rangeFunction as well, similar to blockedAsyncIORequests?
|
|
||
| notFoundErrors: { | ||
| name: 'Not found errors per second', | ||
| type: 'counter', |
There was a problem hiding this comment.
Does this counter signal need a rangeFunction as well, similar to blockedAsyncIORequests?
| // Data transfer signals | ||
| bytesSent: { | ||
| name: 'Bytes sent per second', | ||
| type: 'counter', |
There was a problem hiding this comment.
Does this counter signal need a rangeFunction as well, similar to blockedAsyncIORequests?
| unit: 'percent', | ||
| sources: { | ||
| prometheus: { | ||
| expr: 'increase(windows_iis_server_metadata_cache_hits_total{%(queriesSelector)s}[$__interval:]) / clamp_min(increase(windows_iis_server_metadata_cache_queries_total{%(queriesSelector)s}[$__interval:]),1) * 100', |
There was a problem hiding this comment.
Same comment about [$__interval:], it should be [$__interval:] offset $__interval
| unit: 'percent', | ||
| sources: { | ||
| prometheus: { | ||
| expr: 'increase(windows_iis_server_output_cache_hits_total{%(queriesSelector)s}[$__interval:]) / clamp_min(increase(windows_iis_server_output_cache_queries_total{%(queriesSelector)s}[$__interval:]), 1) * 100', |
There was a problem hiding this comment.
Same comment about [$__interval:], it should be [$__interval:] offset $__interval
| panels.uriCacheHitRatio { gridPos+: { w: 12 } }, | ||
| panels.metadataCacheHitRatio { gridPos+: { w: 12 } }, | ||
| panels.outputCacheHitRatio { gridPos+: { w: 12 } }, | ||
| ]), |
| + g.panel.timeSeries.standardOptions.withUnit('requests') | ||
| + g.panel.timeSeries.standardOptions.withMin(0) | ||
| + g.panel.timeSeries.options.legend.withPlacement('bottom') | ||
| + g.panel.timeSeries.options.tooltip.withMode('none'), |
There was a problem hiding this comment.
Should this be tooltip.withMode('multi') instead?
| ] | ||
| ) | ||
| + g.panel.timeSeries.panelOptions.withDescription('The number of current connections to an IIS site.') | ||
| + g.panel.timeSeries.options.legend.withPlacement('bottom'), |
There was a problem hiding this comment.
Should this panel also define a withMin(0) similar to the blockedAsyncIORequests panel?
3 participants
Screenshots:
Microsoft IIS overview:
Microsoft IIS applications:
