Community. Events v2 API has some limitations not previosuly existing with Events API v1. I'm interested to learn if everyone is aware of this and if it's impacted you negatively?
This is the change:
The entity selector, defining a set of Dynatrace entities to be associated with the event.
Only entities that have been active within the last 24 hours can be selected.
If not set, the event is associated with the environment (dt.entity.environment) entity.
In other words, if your entity (custom or not) has not been active within 24 hours it CAN NOT receive an Events API ingest. Worse yet the JSON response has no helpful information and gives back a 201. Most tools see the 201 and assume all is good, when in fact, it's not good. The Event will never get created.
This is more likely noticed against Custom Entities, Synthetics, or Third-Party Synthetics that act as placeholders for monitoring but are not updated every 1 minute like native entities are.
Looking for some community input on this? Has it impacted you? Have you noticed it? Would this limit your usefulness of Dynatrace?
Solved! Go to Solution.
Just ran into this. This is really a bad decision, for example you cannot push data from external tools to hosts, which are unmonitored (oneagent is disabled) for example.
@wolfgang_beer are there some intentions to improve this? Ideally not limit that at all or limit it to 72 hours or so.
We will introduce an improvement within the next 4 weeks that allows you to target entity id given entities with unlimited time into the past. Eg.: if you have the host ID, you can send an event to that host for years after the host was last seen.
Thank you, really looking forward to this solution. Thanks for hearing us and planning an early fix. Also, making this unlimited time completes another enhancement we were looking for, not sure i made and RFE for it.
Once this is in place it opens so many more opportunities. The next, and i believe final hurdle, will be to fix the Problems features to allow a user to configure longer threshold times. For example Alert if XXXX happens 1 out of every 24 hours. Where the span of time could be up to say 3 months. I'll create and RFE with more details, but once this happens......WOW will we have so much more for Dynatrace to help monitor, it would be amazing. It would be a turning point for which Dynatrace would become fully integrated into our Enterprise as the go-to Alerting and Notification tool.
Somewhat related, we have an already long running case which involves Third-Party Synthetics, that have inconsistent outcomes when querying for application relation, especially with applications that have no RUM and only the 3rd party synth.
We recently have been told that the reason why we sometimes don't see "monitors" (Applications assigned to third party Synthetic Monitoring) is that the timestamp of when the relationship has been last written (between Synthetic monitoring and application) and the timestamp of when the related entity (application) has been last seen needs to be in the query timeframe.
But since the applications do not get (RUM) updates, the timestamps are older.
This causes unpredictable behaviour when quering the linked application.
Also we can not query in the API all the 3rd party entities that have or not have applications related.
So not all all monitors have the relation: