We are using synthetics to monitor various applications, mostly TCP and HTTP plugins are being used to monitor Application and Web Services. We are trying to create a wrapper dashboard where it will show the status of all the monitors on a single visualizaton. For this we have tried using TSVB and Lens Metrics to show the number of monitors showing status as down, and we tried fetching only the last event for each monitor in the visualizaton however, we are struggling to do that. We tried Last Value option available under TSVB however it's not working as intended. Can someone guide how can we pull only the latest event for each monitor and show the count of it based on application of filter monitor.status: down.
Hi Stephen,
This is exactly what we are trying to do, all monitors have a polling interval of 5 mins. We have three visualisations one showing the total number of monitors like 10, and the other two showing how many of these are up and how many of these are down, like 8 up and 2 down. However, there is almost always a mismatch as during the dashboard refresh all visualisations do not refresh simultaneously. The count thus could show like 8 up and 3 down if 1 additional service happens to go down, and would only come into sync after a few additional refreshes have been made.
Hi,
I have three visualizations with the below logic
To find total of errors for a particular application:
cardinality of monitor.id
monitor.status: down and tags: appname
To find Total Application:
cardinality of monitor.id
tags:appname and tags: type of app
To find unhealthy applications:
cardinality of monitor.id
monitor.status: down and tags: app name and tags: app type.
However I am using synthetics with fleet server policy which contains two servers. When I try to visualise these, it does not provide me the correct value all the time. Sometimes it shows and some times it does not show. If I refresh the dashboard the values keeps on changing automatically as if it is flapping.
Also, i would like to know what logic is applied to the synthesis default monitors which shows it as red or green?
Unfortunately it's not easy to replicate it via simple lens viz, though you can try adding summary.final_attempt: true filter and also build viz per location, since monitors running from multiple locations, will not reflect true cardinality.
Another trick you can try is adding monitor.timespan range filter , essentially something like monitor.timespan: now - 15m to now to match only latest run of the document.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.