Uploaded image for project: 'Subscription Watch'
  1. Subscription Watch
  2. SWATCH-2009

tally_measurement values are sometimes getting written to the DB off by 0.000000000000001

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Done
    • Icon: Undefined Undefined
    • 2024-01-03 - API
    • None
    • None

      Sometimes, but not always, the value for tally_measurement.value is getting written to the DB off by 0.000000000000001

      For example, an hourly tally produced a value of 9.799999999999999 initially. Then after clearing the DB of everything except events, and running the tallying again, it ended up as 9.8

      I have also seen this in the opposite direction where the expected value is 9.6 but results in 9.600000000000001

       

      NOTE: This issue appears to be random and is not always reproducible. Sometimes not including the last event below seems to reproduce more frequently.

      Steps To Reproduce:
      1. Load some events into the DB

      INSERT INTO public.events (event_id, "timestamp", data, event_type, event_source, instance_id, org_id, metering_batch_id, record_date) VALUES ('ac69e237-1387-4236-8099-33ef204210d8', '2023-12-05 07:00:00-04', '{"sla": "Premium", "role": "osd", "org_id": "org123", "event_id": "ac69e237-1387-4236-8099-33ef204210d8", "timestamp": "2023-12-05T11:00:00+00:00", "event_type": "snapshot_redhat.com:openshift_dedicated:4cpu_hour", "expiration": "2023-12-05T12:00:00+00:00", "instance_id": "6137ea0b-7d79-4caa-9c15-9fb2470876cd", "display_name": "6137ea0b-7d79-4caa-9c15-9fb2470876cd", "event_source": "prometheus", "measurements": [{"uom": "Cores", "value": 2.9}], "product_tag": ["rosa"], "service_type": "OpenShift Cluster", "account_number": "account_org123", "billing_provider": "red hat", "billing_account_id": "redhat1"}', 'snapshot_redhat.com:openshift_dedicated:4cpu_hour', 'prometheus', '6137ea0b-7d79-4caa-9c15-9fb2470876cd', 'org123', NULL, '2023-12-06 14:16:55.418313-04');
INSERT INTO public.events (event_id, "timestamp", data, event_type, event_source, instance_id, org_id, metering_batch_id, record_date) VALUES ('3dea8bdb-fd1c-43c8-8cf7-414d43a514bf', '2023-12-05 07:00:00-04', '{"sla": "Premium", "role": "osd", "org_id": "org123", "event_id": "3dea8bdb-fd1c-43c8-8cf7-414d43a514bf", "timestamp": "2023-12-05T11:00:00+00:00", "event_type": "snapshot_redhat.com:openshift_dedicated:4cpu_hour", "expiration": "2023-12-05T12:00:00+00:00", "instance_id": "fe304292-a71d-4d2d-a1ea-02095dde5d04", "display_name": "fe304292-a71d-4d2d-a1ea-02095dde5d04", "event_source": "prometheus", "measurements": [{"uom": "Cores", "value": 3.0}], "product_tag": ["rosa"], "service_type": "OpenShift Cluster", "account_number": "account_org123", "billing_provider": "red hat", "billing_account_id": "redhat1"}', 'snapshot_redhat.com:openshift_dedicated:4cpu_hour', 'prometheus', 'fe304292-a71d-4d2d-a1ea-02095dde5d04', 'org123', NULL, '2023-12-06 14:16:55.418607-04');
INSERT INTO public.events (event_id, "timestamp", data, event_type, event_source, instance_id, org_id, metering_batch_id, record_date) VALUES ('98c8252f-57d8-456d-b51a-6ca9f381d6b3', '2023-12-05 07:00:00-04', '{"sla": "Premium", "role": "osd", "org_id": "org123", "event_id": "98c8252f-57d8-456d-b51a-6ca9f381d6b3", "timestamp": "2023-12-05T11:00:00+00:00", "event_type": "snapshot_redhat.com:openshift_dedicated:4cpu_hour", "expiration": "2023-12-05T12:00:00+00:00", "instance_id": "ac519a5a-bc56-45ac-b1e6-32a3140d9f2d", "display_name": "ac519a5a-bc56-45ac-b1e6-32a3140d9f2d", "event_source": "prometheus", "measurements": [{"uom": "Cores", "value": 2.3}], "product_tag": ["rosa"], "service_type": "OpenShift Cluster", "account_number": "account_org123", "billing_provider": "red hat", "billing_account_id": "redhat1"}', 'snapshot_redhat.com:openshift_dedicated:4cpu_hour', 'prometheus', 'ac519a5a-bc56-45ac-b1e6-32a3140d9f2d', 'org123', NULL, '2023-12-06 14:16:55.418753-04');
INSERT INTO public.events (event_id, "timestamp", data, event_type, event_source, instance_id, org_id, metering_batch_id, record_date) VALUES ('97a860be-f52d-46db-838c-414b3ee82fe9', '2023-12-05 07:00:00-04', '{"sla": "Premium", "role": "osd", "org_id": "org123", "event_id": "97a860be-f52d-46db-838c-414b3ee82fe9", "timestamp": "2023-12-05T11:00:00+00:00", "event_type": "snapshot_redhat.com:openshift_dedicated:4cpu_hour", "expiration": "2023-12-05T12:00:00+00:00", "instance_id": "7681cdcf-e2c9-4532-bfe0-83b38374c116", "display_name": "7681cdcf-e2c9-4532-bfe0-83b38374c116", "event_source": "prometheus", "measurements": [{"uom": "Cores", "value": 1.4}], "product_tag": ["rosa"], "service_type": "OpenShift Cluster", "account_number": "account_org123", "billing_provider": "red hat", "billing_account_id": "redhat1"}', 'snapshot_redhat.com:openshift_dedicated:4cpu_hour', 'prometheus', '7681cdcf-e2c9-4532-bfe0-83b38374c116', 'org123', NULL, '2023-12-06 14:16:55.418892-04');
INSERT INTO public.events (event_id, "timestamp", data, event_type, event_source, instance_id, org_id, metering_batch_id, record_date) VALUES ('753366fc-3591-4ebc-97f2-0f58c1b4d6cb', '2023-12-05 07:00:00-04', '{"sla": "Premium", "role": "osd", "org_id": "org123", "event_id": "753366fc-3591-4ebc-97f2-0f58c1b4d6cb", "timestamp": "2023-12-05T11:00:00+00:00", "event_type": "snapshot_redhat.com:openshift_dedicated:4cpu_hour", "expiration": "2023-12-05T12:00:00+00:00", "instance_id": "e3a5225b-a826-43ab-9374-0c4fe40fde60", "display_name": "e3a5225b-a826-43ab-9374-0c4fe40fde60", "event_source": "prometheus", "measurements": [{"uom": "Cores", "value": 0.2}], "product_tag": ["rosa"], "service_type": "OpenShift Cluster", "account_number": "account_org123", "billing_provider": "red hat", "billing_account_id": "redhat1"}', 'snapshot_redhat.com:openshift_dedicated:4cpu_hour', 'prometheus', 'e3a5225b-a826-43ab-9374-0c4fe40fde60', 'org123', NULL, '2023-12-06 14:16:55.419027-04');

      2. Run the hourly tally.

      http POST ":8000/api/rhsm-subscriptions/v1/internal/tally/hourly?org=org123&start=2023-12-05T00:00Z&end=2023-12-06T23:00Z" x-rh-swatch-psk:placeholder Origin:console.redhat.com

      3. Check the current values from the tally_measurements table

      psql -U rhsm-subscriptions rhsm-subscriptions -c "select value from tally_measurements"

      4. Clear the data.

      psql -U rhsm-subscriptions rhsm-subscriptions -c "truncate hosts cascade; truncate tally_snapshots cascade; truncate billable_usage_remittance cascade;"

      5. Re-run the hourly tally.
      6. Check the current values from the tally_measurements table

      psql -U rhsm-subscriptions rhsm-subscriptions -c "select value from tally_measurements"

      Done

      • Find out why this is happening
      • Share with the team why it's happening
      • Make it stop happening (or create new cards to be refined if the solution is complex)

              jcarvaja@redhat.com Jose Carvajal Hilario
              mstead@redhat.com Michael Stead
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                Resolved: