General - Catch me up to speed - Tagging - Segments - Openpipeline - Metrics - Logs

[Mullaneyb]

Been using Dynatrace for a very long time but am now leaning into use cases I have not approached before.

Just after a sanity check 😀

Take Azure data factory logs and alert the respective team on a list of given scenarios

Workflow is as so:

1. Log capture (openpipeline) - match the relevant conditions
2. Processing (openpipeline) - Parse relevant JSON fields into their own columns, some are nested, also extract specific messages discarding irrelevant content.
3. Metric Extraction (openpipeline) - add the columns as metric dimensions i.e azure.resource.name, dt.source_entity, pipelineName
4. Metric Event(classic) - set the metric event with placeholders etc. ALSO set metric property dt.source_entity: {dims:dt.source_entity} to scope the metric event to the log source entity instead of the Entire Dynatrace Environment. 
5. Alerting Profile (classic) - setup a profile simple filter of event starts with "xxyy"
6. Email Integration (Classic) - set up a problem integration in this case Email.

The requirement I have is to have the {dims} object available in the Subject line of the Email Integration, but this doesn't seem possible? the Metric event has access to the properties and can be used to supplement the email body with {dims:value} etc via the {ProblemDetailsHTML} object. Dimensions don't seem directly accessible from the Email Integration. If possible, I would like to have this dynamic otherwise I will be looking at many many many rules.

EDIT / Solution:
Set Email Subject to {ImpactedEntity}
Set Event template | Title to the dims pattern you want i.e {dims:azure.resource.name} - {dims:MessageInner}

Tagging etcetera. 

I came from a tag the host and tag the service then propagate inwards to PG era. I'm more learning towards tag PG (i.e one rule) propagate to host and service. Seems to work fine except when certain PG's (window services etc) are grouped across the entire platform and therefore have every app tag. Historically I would have had the PG split in my mind it's a different process group if it's a different application. I understand host groups would solve my issue here but id usually use that for env not app.

Most of the documentation currently seems towards tagging generically but providing that data from the agent install point (too late for me) 

When it comes to Tagging Metrics, Event, Logs, for something like use in Segments but in the way they are all joined to the collective application host/pg/service tag what's the approach.

Example might be how do I tag every single entity/entity type that belongs to an application where I can split or filter by environment. use case would be dashboard templates. make it once, just filter on segment APP + ENV. whereas in classic land I would copy the dashboard and replace all the values, or in some cases Management zone filtering. 

I understand Dynatrace is in a transition phase, and I might be trying to force the old into the new (or vise versa) and its just incompatible. 

[Mullaneyb]

Additionally, I have a use case for DQL where I need to apply some calculated fields to logs to pertain if a service is up down or degraded. I'm having trouble thinking of how this will work. As below 

I tried something like the below but I need the 5 scenarios handled. A IF ELSEIF would be nice? as if I set the Calculation column again with another set of logic it overwrites it

| fieldsAdd Calculation = if(column5 == 1 and column6 == 1) or (column3 == 1 and column2 == 1), 1,Else: 0)
| fieldsAdd Calculation = if(column1 == 1 and column4 == 1) , 2,Else: 0)
| fieldsAdd Calculation = if(column2 == 0 and column3 == 0) or (column5 == 0 and column6 == 0), 1,Else: 0)

[Mullaneyb]

Additionally. How can one propagate host tags to both LOGS and METRICS for a variable filter in segments?

I have no issue choosing the application and environment for the other data types but am struggling to see how to do it for these two types as there is no metadata for it?