Raise a problem when license consumption is close to limit

[deni]

Hi,

We are using a Dynatrace Managed environment and would like to automatically raise a problem when license consumption (for example, DDUs) is close to being exhausted.

I checked the available metrics but couldn’t find anything like ddu.used or hostunits.used. I did see some metrics such as billing.ddu.log.total, but for DDUs there are also metrics and traces that need to be considered in the calculation, not just logs.

• Is there a way to monitor the overall license consumption across all DDU categories?

• Can we set up alerting or problem generation for approaching license limits in Dynatrace Managed?

• If there are no direct metrics, what would be the recommended approach to achieve this?

Any advice or best practices would be greatly appreciated.

Regards, Deni

[AntonPineiro]

Hi,

Managed is more limited about that.

For DDU categories, you can try this dashboard.

I do not know any way to alert automatically in Managed about that. I would say review before dashboard to checking trend, and compare to limit your license has.-

Best regards

[Julius_Loman]

For these purposes we developed an Extension 2.0 which creates metrics per Dynatrace environment and you can then alert using Metric events. It scrapes the Cluster API and writes metrics. 

[deni]

@Julius_LomanThanks for the reply!

If I understand correct you suggest to create a custom extension which calls API to collect the license information and the sent a custom event via v2/events/ingest?

I searched in the swagger but can't find API to get licenses - I see only v1/config/clusterid  and v1/clusterversion nothing more in v1, v2 and config APIs connected to cluster or license.

[Julius_Loman]

@deni  - https://docs.dynatrace.com/managed/shortlink/api-cluster-environments-list-existing
This includes consumption data (Classic licensing per environment).

Mark's answer is also valid if you need it for a single environment. We had a requirement to alert on DDU consumption for two environments together, which is not possible with metrics just from a single environment.

[mark_bley · Accepted solution]

In managed you should know from your license the amount of DDUs you have (i.e. a fixed number)

What you can do is have a metric event to alert on this, but keep in mind this will be per environment.

((builtin:billing.ddu.metrics.total:splitBy():sum:default(0):fold(sum))
+
(builtin:billing.ddu.log.total:splitBy():sum:default(0):fold(sum))
+
(builtin:billing.ddu.serverless.total:splitBy():sum:default(0):fold(sum))
+
(builtin:billing.ddu.events.total:splitBy():sum:default(0):fold(sum))
+
(builtin:billing.ddu.traces.total:splitBy():sum:default(0):fold(sum)))

 That will give you the total amount used.

Let's say your total available is 1M, so put as a static threshold in the metric event with something like 800k.

With an extension, script or synthetic-test you could extract this information for the whole cluster from this API https://docs.dynatrace.com/managed/shortlink/api-cluster-get-license-consumption

Also maybe something to have a look at are DDU pools, where you can assign a specific env a limit of DDUs allowed to be consumed. https://docs.dynatrace.com/docs/shortlink/davis-data-units#ddu-pools

[deni]

Hi,

I'm pasting the code which I'm using:

with requests.get(f"{DT_CLUSTER_URL}/api/cluster/v2/license/consumption", headers=headers, stream=True, verify=VERIFY_SSL) as response:
...
with inner_zip.open(json_filename) as json_file:
                with io.TextIOWrapper(json_file, encoding="utf-8") as text_file:
                    data = json.load(text_file)
                    env_entries = data.get("environmentBillingEntries", [])

                    for env in env_entries:
                        # Sum DDUs
                        for pool in env.get("davisDataUnits", []):
                            total_ddus += safe_float(pool.get("total", 0))

                        # Count hosts
                        host_usages = env.get("hostUsages", [])
                        total_hostunits += len(host_usages)

                        # Count DEMs (based on agent usages per host)
                        for host in host_usages:
                            total_dems += len(host.get("agentUsages", []))
print("Aggregated License Consumption (from license start until now):")
print(f"  Total DDUs used: {total_ddus:.2f}")
print(f"  Total DEMs used: {total_dems}")
print(f"  Total Host Units used: {total_hostunits}")

I saw that there are two environments in this Managed Dynatrace Cluster. The information on the /cmc shows aggregating result for the current consumed/available licenses and that's why I'm also aggregating the results. The problem is that the results which I get completely differs from these shown in the /cmc:

Aggregated License Consumption (from license start until now):
  Total DDUs used: 93747.71
  Total DEMs used: 4428
  Total Host Units used: 4424

but in the /cmc it is:

Host Units: 15/15
DEM: 2.07k/10k
DDU: 599.51k/900k

Can you help me with that?

Regards, Deni