https://community.databricks.com/t5/technical-blog/triggering-databricks-jobs-from-celonis-action-flows-a-step-by/ba-p/156263 <P>Turn process intelligence into automated action: connect Celonis directly to Databricks.</P> <H2><SPAN>Why Connect Celonis Action Flows to Databricks?</SPAN></H2> <P><SPAN>Celonis shows you where your business processes break down — where they stall, deviate, and leak money. But seeing the problem isn't the same as fixing it.</SPAN></P> <P><SPAN>Databricks is where your data pipelines and ML models run. By connecting Celonis Action Flows directly to the Databricks Jobs API, you turn process insights into automated data actions — no middleware, no webhook servers, no glue code. One HTTP call.</SPAN></P> <H2><SPAN>Use Cases</SPAN></H2> <H3><SPAN>1. Late Delivery Prediction &amp; Mitigation</SPAN></H3> <P><SPAN>Celonis detects that a customer order has been in "warehouse processing" for longer than the 95th percentile. The Action Flow triggers a Databricks job that scores the order against a delivery delay prediction model, checks alternative fulfillment options in the lakehouse, and writes a recommended action back to the order management system.</SPAN></P> <P><STRONG>Impact:</STRONG><SPAN> Orders at risk of late delivery are flagged and rerouted before the customer notices — not after the SLA is breached.</SPAN></P> <H3><SPAN>2. Automated Compliance Alerting</SPAN></H3> <P><SPAN>Celonis flags a process execution that violates a segregation-of-duties policy — the same person created and approved a purchase order. The Action Flow triggers a Databricks job that logs the violation with full event context into an audit Delta table, runs a compliance scoring model across all recent transactions, and updates the compliance dashboard.</SPAN></P> <P><STRONG>Impact:</STRONG><SPAN> Violations are caught and documented in real time, with a full audit trail, not discovered during the quarterly review.</SPAN></P> <H3><SPAN>3. Process-Aware Model Retraining</SPAN></H3> <P><SPAN>Celonis identifies that a process variant has shifted — a new pattern is emerging that the current prediction models don't account for. The Action Flow triggers a Databricks ML pipeline that extracts the latest process event data, retrains the model with the new variant, validates performance, and deploys the updated model to the serving endpoint.</SPAN></P> <P><STRONG>Impact:</STRONG><SPAN> ML models stay aligned with how the business actually operates today, not how it operated when the model was last trained.</SPAN></P> <H2><SPAN>Architecture</SPAN></H2> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_0-1781271627643.png" style="width: 687px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/27748i285775E5FF0F63DD/image-dimensions/687x341?v=v2" width="687" height="341" role="button" title="TarziSimon_0-1781271627643.png" alt="TarziSimon_0-1781271627643.png" /></span></P> <P><SPAN>No middleware. Celonis calls Databricks directly over HTTPS.</SPAN></P> <H2><SPAN>Prerequisites</SPAN></H2> <P><SPAN>Before you start:</SPAN></P> <UL> <LI style="font-weight: 400;" aria-level="1"><SPAN>A </SPAN><STRONG>Celonis</STRONG><SPAN> account with access to Action Flows</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>A </SPAN><STRONG>Databricks</STRONG><SPAN> workspace with at least one job configured</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>A </SPAN><STRONG>Databricks Personal Access Token (PAT) for the easier setup (for production OAuth M2M is also supported)&nbsp;</STRONG></LI> </UL> <H3><SPAN>Create a Databricks PAT</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>In Databricks, click your </SPAN><STRONG>username</STRONG><SPAN> (top-right)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Settings</STRONG><SPAN> &gt; </SPAN><STRONG>Developer</STRONG><SPAN> &gt; </SPAN><STRONG>Access tokens</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Manage</STRONG><SPAN> &gt; </SPAN><STRONG>Generate new token</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Name it </SPAN><SPAN>celonis-action-flow</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set an expiry (90 days recommended — rotate regularly)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Generate</STRONG><SPAN> and </SPAN><STRONG>copy the token immediately</STRONG></LI> </OL> <H3><SPAN>Get Your Job ID</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>In Databricks, go to </SPAN><STRONG>Workflows</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click the job you want to trigger</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>The </SPAN><STRONG>Job ID</STRONG><SPAN> is in the URL: </SPAN><SPAN>https://&lt;workspace&gt;/#job/&lt;job-id&gt;</SPAN></LI> </OL> <H2><SPAN>Step-by-Step Tutorial</SPAN></H2> <H3><SPAN>Step 1: Create an Action Flow</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Log into Celonis</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Navigate to </SPAN><STRONG>Action Flows</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Create Action Flow</STRONG></LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_1-1778074571012.png" style="width: 400px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26686i55A286DC626B4F67/image-size/medium?v=v2&amp;px=400" role="button" title="TarziSimon_1-1778074571012.png" alt="TarziSimon_1-1778074571012.png" /></span></P> <H3><SPAN>Step 2: Configure the Trigger</SPAN></H3> <P><SPAN>Click the trigger module and choose what starts the flow:</SPAN></P> <UL> <LI style="font-weight: 400;" aria-level="1"><STRONG>Schedule</STRONG><SPAN> — run on a timer (e.g., every hour)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Celonis Signal</STRONG><SPAN> — triggered when a process condition is met (e.g., order stuck &gt; 48 hours)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Manual</STRONG><SPAN> — triggered by a user clicking a button</SPAN></LI> </UL> <P><SPAN>For this tutorial, choose </SPAN><STRONG>Schedule</STRONG><SPAN> and set it to every hour.</SPAN></P> <H3><SPAN>Step 3: Add an HTTP Module</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>+</STRONG><SPAN> after the trigger</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Search for </SPAN><STRONG>HTTP</STRONG><SPAN> and select </SPAN><STRONG>Make a Request</STRONG></LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_2-1778074571012.png" style="width: 513px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26687i45F25BAD821B4FC0/image-dimensions/513x316?v=v2" width="513" height="316" role="button" title="TarziSimon_2-1778074571012.png" alt="TarziSimon_2-1778074571012.png" /></span></P> <H3><SPAN>Step 4: Configure the Databricks API Call</SPAN></H3> <P><SPAN>Fill in the HTTP module:</SPAN></P> <P><STRONG>URL:<BR /></STRONG><SPAN>AWS: <A href="https://&lt;your-workspace&gt;.cloud.databricks.com/api/2.1/jobs/run-now" target="_blank" rel="noopener">https://&lt;your-workspace&gt;.cloud.databricks.com/api/2.1/jobs/run-now</A><BR /></SPAN><SPAN>Azure: <A href="https://adb-&lt;workspace-id&gt;.&lt;shard&gt;.azuredatabricks.net/api/2.1/jobs/run-now" target="_blank" rel="noopener">https://adb-&lt;workspace-id&gt;.&lt;shard&gt;.azuredatabricks.net/api/2.1/jobs/run-now</A><BR /></SPAN><SPAN>GCP: <A href="https://&lt;workspace-id&gt;.gcp.databricks.com/api/2.1/jobs/run-now" target="_blank" rel="noopener">https://&lt;workspace-id&gt;.gcp.databricks.com/api/2.1/jobs/run-now</A><BR /></SPAN></P> <P><STRONG>Method:</STRONG><SPAN> POST</SPAN></P> <P><STRONG>Headers:</STRONG></P> <TABLE> <TBODY> <TR> <TD> <P><STRONG>Key</STRONG></P> </TD> <TD> <P><STRONG>Value</STRONG></P> </TD> </TR> <TR> <TD> <P><SPAN>Authorization</SPAN></P> </TD> <TD> <P><SPAN>Bearer dapi... (your PAT)</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>Content-Type</SPAN></P> </TD> <TD> <P><SPAN>application/json</SPAN></P> </TD> </TR> </TBODY> </TABLE> <P><STRONG>Body type:</STRONG><SPAN> Raw (JSON)<BR /></SPAN><STRONG>Body:<BR /></STRONG></P> <LI-CODE lang="markup">{ "job_id": 12345 }</LI-CODE> <P><SPAN>Replace the workspace URL and job ID with your actual values.<BR /></SPAN><SPAN>Set the “</SPAN><SPAN>Parse response” to Yes, so we can refer to the job id in later steps.&nbsp;</SPAN></P> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_3-1778074571013.png" style="width: 513px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26689i4E798784FA95FBE2/image-dimensions/513x392?v=v2" width="513" height="392" role="button" title="TarziSimon_3-1778074571013.png" alt="TarziSimon_3-1778074571013.png" /></span></P> <H3><SPAN>Step 5: Pass Dynamic Parameters (Optional)</SPAN></H3> <P><SPAN>To send data from Celonis into the Databricks job — for example, a case ID or threshold value:</SPAN></P> <LI-CODE lang="markup">{ "job_id": 12345, "notebook_params": { "case_id": "{{celonis.case_id}}", "event_type": "{{celonis.signal_name}}", "threshold_days": "5" } }</LI-CODE> <P><SPAN>In your Databricks notebook, read these with:</SPAN></P> <LI-CODE lang="markup">case_id = dbutils.widgets.get("case_id") event_type = dbutils.widgets.get("event_type")</LI-CODE> <H3><SPAN>Step 6: Test the Request</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Run once</STRONG><SPAN> in the Action Flow editor</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>You should get a </SPAN><STRONG>200 OK</STRONG><SPAN> response:</SPAN></LI> </OL> <LI-CODE lang="markup">{ "run_id": 67890, "number_in_job": 1 }</LI-CODE> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_4-1778074571013.png" style="width: 510px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26690i811806A657D99BF0/image-dimensions/510x390?v=v2" width="510" height="390" role="button" title="TarziSimon_4-1778074571013.png" alt="TarziSimon_4-1778074571013.png" /></span></P> <H3><SPAN>Step 7: Verify in Databricks</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Go to </SPAN><STRONG>Workflows</STRONG><SPAN> in Databricks</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click your job</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>You should see a new run triggered by the API call</SPAN></LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_5-1778074571013.png" style="width: 570px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26691iF3BE0546A30956D4/image-dimensions/570x436?v=v2" width="570" height="436" role="button" title="TarziSimon_5-1778074571013.png" alt="TarziSimon_5-1778074571013.png" /></span></P> <P>&nbsp;</P> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_6-1778074571014.png" style="width: 573px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26692iA77F9241CE136975/image-dimensions/573x439?v=v2" width="573" height="439" role="button" title="TarziSimon_6-1778074571014.png" alt="TarziSimon_6-1778074571014.png" /></span></P> <P>&nbsp;</P> <H3><SPAN>Step 8: Add Error Handling</SPAN></H3> <P><SPAN>Add a </SPAN><STRONG>Router</STRONG><SPAN> after the HTTP module to split the flow into success and failure paths.</SPAN></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>+</STRONG><SPAN> after the HTTP module</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Add a </SPAN><STRONG>Router</STRONG><SPAN> module — this splits the flow into two branches</SPAN></LI> </OL> <P><STRONG>Configure the Success path:</STRONG></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click the dotted line on the first path</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Set up a filter</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Label: </SPAN><SPAN>Success</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Condition: set it to </SPAN><STRONG>Status code</STRONG> <SPAN>Equal to</SPAN> <SPAN>200</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>After this filter, add your next module (e.g., the polling step from Step 9, or a log entry)</SPAN></LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_7-1778074571014.png" style="width: 549px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26693i883488DABB587B96/image-dimensions/549x420?v=v2" width="549" height="420" role="button" title="TarziSimon_7-1778074571014.png" alt="TarziSimon_7-1778074571014.png" /></span></P> <P>&nbsp;</P> <P><STRONG>Configure the Failure path:</STRONG></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click the dotted line on the second path</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Set up a filter</STRONG></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Label: </SPAN><SPAN>Failure</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Condition: set it to </SPAN><STRONG>Status code</STRONG> <SPAN>Not equal to</SPAN> <SPAN>200</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>After this filter, add an alert module:</SPAN></LI> </OL> <UL> <LI style="font-weight: 400;" aria-level="1"><STRONG>Email</STRONG><SPAN> — send yourself a failure notification</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Slack</STRONG><SPAN> — post an alert to a channel</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Or a Text Aggregator</STRONG><SPAN> — log the error for debugging</SPAN></LI> </UL> <P><STRONG>Where to find "Status code"</STRONG><SPAN> in the filter condition: click the value field and you will see the output variables from the previous HTTP module. Look for the </SPAN><STRONG>Status code</STRONG><SPAN> field — it is a number like 200, 400, 403, etc.</SPAN></P> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_8-1778074571014.png" style="width: 501px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26694iE35A946898B9E76D/image-dimensions/501x383?v=v2" width="501" height="383" role="button" title="TarziSimon_8-1778074571014.png" alt="TarziSimon_8-1778074571014.png" /></span></P> <H3><SPAN>Step 9: Poll for Job Completion (Optional)</SPAN></H3> <P><SPAN>If the Action Flow needs to wait for the Databricks job to finish before continuing:</SPAN></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Add a </SPAN><STRONG>Sleep</STRONG><SPAN> module — wait 30 seconds</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Add another </SPAN><STRONG>HTTP</STRONG><SPAN> module:</SPAN></LI> </OL> <P><SPAN>AWS: GET <A href="https://&lt;workspace-id&gt;.cloud.databricks.com/api/2.1/jobs/runs/get?run_id={{run_id}}" target="_blank" rel="noopener">https://&lt;workspace-id&gt;.cloud.databricks.com/api/2.1/jobs/runs/get?run_id={{run_id}}</A><BR /></SPAN><SPAN>Azure: GET <A href="https://adb-" target="_blank" rel="noopener">https://adb-</A>&lt;workspace-id&gt;.&lt;shard&gt;.azuredatabricks.net/api/2.1/jobs/runs/get?run_id={{run_id}}<BR /></SPAN><SPAN>GCP: GET https://&lt;workspace-id&gt;.gcp.databricks.com/api/2.1/jobs/runs/get?run_id={{run_id}}</SPAN></P> <P><SPAN>Where </SPAN><SPAN>{{run_id}}</SPAN><SPAN> = click the field, pick</SPAN> <SPAN>run_id</SPAN> <SPAN>from the first HTTP module's output. If you don't see a parsed body, the response might not be parsed as JSON. Check that the first HTTP module has </SPAN><STRONG>Parse response</STRONG><SPAN> set to </SPAN><STRONG>Yes</STRONG><SPAN>, and run the Flow once at least so Celonis can parse the output once before referring to it.&nbsp;</SPAN></P> <P><SPAN>Check the response:</SPAN></P> <P><SPAN>After the second HTTP module (the GET status check), add a </SPAN><STRONG>Router</STRONG><SPAN> with two paths:</SPAN></P> <P><STRONG>Path 1 — Job finished:</STRONG><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</SPAN></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click the dotted line on the first path</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set up a filter</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Label: </SPAN><SPAN>Job Done</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Condition: </SPAN><STRONG>data.state.life_cycle_state</STRONG> <SPAN>Equal to TERMINATED</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set up notification&nbsp;</SPAN></LI> </OL> <P><SPAN>&nbsp;&nbsp;</SPAN><STRONG>Path 2 — Job still running:</STRONG></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click the dotted line on the second path</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set up a filter</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Label:</SPAN> <SPAN>Still Running</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Condition: </SPAN><STRONG>data.state.life_cycle_state</STRONG> <SPAN>Not equal to TERMINATED</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set up notifications</SPAN>&nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;&nbsp;</LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_9-1778074571014.png" style="width: 523px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26695iFBB9F0E60B28EA72/image-dimensions/523x400?v=v2" width="523" height="400" role="button" title="TarziSimon_9-1778074571014.png" alt="TarziSimon_9-1778074571014.png" /></span></P> <P><SPAN>Or&nbsp;</SPAN></P> <P><SPAN>Use a </SPAN><STRONG>Repeater</STRONG><SPAN> module instead of a loop. Place it right after the first Router's success path:&nbsp;</SPAN></P> <P><STRONG>Flow structure:</STRONG><SPAN>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</SPAN></P> <P><SPAN>&nbsp;&nbsp;Trigger → HTTP (trigger job) → Router (200 or error) → Repeater (20 iterations) → Sleep 30s → HTTP (check status) → Router → "Job Done" → next step → "Still Running" → does nothing, repeater continues</SPAN></P> <P><SPAN>&nbsp;&nbsp;</SPAN><STRONG>Setup:</STRONG></P> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>After the success path of the first Router, add a </SPAN><STRONG>Repeater</STRONG><SPAN> module</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>Set </SPAN><STRONG>Repeats</STRONG><SPAN> to </SPAN><SPAN>20</SPAN><SPAN> (that gives you 20 x 30 seconds = 10 minutes max wait)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>After the Repeater, add a </SPAN><STRONG>Sleep</STRONG><SPAN> module — set to </SPAN><SPAN>30</SPAN><SPAN> seconds</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>After Sleep, add the </SPAN><STRONG>HTTP</STRONG><SPAN> module (GET status check)</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>After the HTTP module, add a </SPAN><STRONG>Router</STRONG><SPAN> with a filter on the first path:&nbsp;<BR /></SPAN><SPAN>5.1. Label: </SPAN><SPAN>Job Done</SPAN><SPAN>&nbsp;<BR /></SPAN><SPAN>5.2. Condition: </SPAN><STRONG>data.state.life_cycle_state</STRONG> <SPAN>Equal to TERMINATED&nbsp;<BR /></SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>On the "Job Done" path, continue with your next steps (or just end the flow)&nbsp;&nbsp;</SPAN></LI> </OL> <H3><SPAN>Step 10: Activate</SPAN></H3> <OL> <LI style="font-weight: 400;" aria-level="1"><SPAN>Click </SPAN><STRONG>Activate</STRONG><SPAN> to turn the Action Flow on</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><SPAN>It now runs automatically based on your trigger</SPAN></LI> </OL> <P><span class="lia-inline-image-display-wrapper lia-image-align-inline" image-alt="TarziSimon_10-1778074571015.png" style="width: 525px;"><img src="https://community.databricks.com/t5/image/serverpage/image-id/26696i7A50E999F926E06D/image-dimensions/525x402?v=v2" width="525" height="402" role="button" title="TarziSimon_10-1778074571015.png" alt="TarziSimon_10-1778074571015.png" /></span></P> <H2><SPAN>Security Best Practices</SPAN></H2> <TABLE> <TBODY> <TR> <TD> <P><STRONG>Practice</STRONG></P> </TD> <TD> <P><STRONG>Why</STRONG></P> </TD> </TR> <TR> <TD> <P><SPAN>Use a **service principal** instead of a personal PAT</SPAN></P> </TD> <TD> <P><SPAN>Service principals are not tied to one user and can be scoped precisely</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>**Store the token** in Celonis connection/secret management</SPAN></P> </TD> <TD> <P><SPAN>Never hardcode tokens in the Action Flow body</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>Set **PAT expiry to 90 days** and rotate</SPAN></P> </TD> <TD> <P><SPAN>Limits the blast radius if the token leaks</SPAN></P> </TD> </TR> <TR> <TD> <P><STRONG>Restrict PAT permissions</STRONG></P> </TD> <TD> <P><SPAN>Only grant access to the specific jobs needed</SPAN></P> </TD> </TR> <TR> <TD> <P><STRONG>IP allowlisting</STRONG></P> </TD> <TD> <P><SPAN>Restrict Databricks API access to Celonis egress IPs</SPAN></P> </TD> </TR> </TBODY> </TABLE> <H2><SPAN><BR />Troubleshooting</SPAN></H2> <TABLE> <TBODY> <TR> <TD> <P><STRONG>Problem</STRONG></P> </TD> <TD> <P><STRONG>Solution</STRONG></P> </TD> </TR> <TR> <TD> <P><SPAN>401 Unauthorized</SPAN></P> </TD> <TD> <P><SPAN>PAT is invalid or expired — generate a new one</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>403 Forbidden</SPAN></P> </TD> <TD> <P><SPAN>PAT doesn't have permission for this job — check job permissions</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>404 Not Found</SPAN></P> </TD> <TD> <P><SPAN>Wrong workspace URL or job ID — double-check both</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>400 Bad Request</SPAN></P> </TD> <TD> <P><SPAN>Malformed JSON — validate syntax (missing comma, wrong quotes)</SPAN></P> </TD> </TR> <TR> <TD> <P><SPAN>Job runs but fails</SPAN></P> </TD> <TD> <P><SPAN>API call succeeded — check the job run logs in Databricks</SPAN></P> </TD> </TR> </TBODY> </TABLE> <H2><SPAN><BR />What's Next</SPAN></H2> <P><SPAN>Once the basic connection works, you can:</SPAN></P> <UL> <LI style="font-weight: 400;" aria-level="1"><STRONG>Chain jobs</STRONG><SPAN> — trigger a sequence of Databricks jobs from a single Action Flow</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Write results back</STRONG><SPAN> — use the Celonis Data Push API to send Databricks output back into Celonis data models</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Call ML endpoints</STRONG><SPAN> — hit a Databricks Model Serving endpoint for real-time predictions instead of batch jobs</SPAN></LI> <LI style="font-weight: 400;" aria-level="1"><STRONG>Build closed loops</STRONG><SPAN> — detect issue in Celonis → analyze in Databricks → push fix to source system → verify in Celonis</SPAN></LI> </UL> <P><SPAN>Process intelligence plus data intelligence. That's the closed loop.</SPAN></P> <P><I><SPAN>For questions or feedback, reach out to your Databricks Account team.</SPAN></I></P> Fri, 12 Jun 2026 14:10:40 GMT Tarzi-Simon 2026-06-12T14:10:40Z https://community.databricks.com/t5/technical-blog/triggering-databricks-jobs-from-celonis-action-flows-a-step-by/ba-p/156263 <P class="wnfdntf _1ibi0s3e6 _1ibi0s3ce _1ibi0s3db" data-pm-slice="1 1 []">Learn how to trigger <STRONG>Databricks Jobs</STRONG> directly from <STRONG>Celonis Action Flows</STRONG> using a simple HTTP call. This step-by-step guide walks through setup, dynamic parameters, error handling, polling, and security best practices to turn process insights into automated action.</P> Fri, 12 Jun 2026 14:10:40 GMT https://community.databricks.com/t5/technical-blog/triggering-databricks-jobs-from-celonis-action-flows-a-step-by/ba-p/156263 Tarzi-Simon 2026-06-12T14:10:40Z