5 things I wish I knew before building Sentinel playbooks in Logic Apps:
1. Use the incident trigger unless you truly need one run per alert
2. Managed identity + Key Vault for API keys. Never hardcode secrets
3. Cache enrichment results (Azure Table Storage works well) so the same IP doesn't burn your API quota
4. Filter private and reserved IPs before calling any external API
5. Always write results back to the incident as a comment, failures included
Automation nobody can see is automation nobody trusts. #SOAR
31
5 things I wish I knew before building Sentinel playbooks in Logic Apps:
1. Use the incident trigger unless you truly need one run per alert
2. Managed identity + Key Vault for API keys. Never hardcode secrets
3. Cache enrichment results (Azure Table Storage works well) so the same IP doesn't burn your API quota
4. Filter private and reserved IPs before calling any external API
5. Always write results back to the incident as a comment, failures included
Automation nobody can see is automation nobody trusts. #SOAR
20
Alert vs Incident in Microsoft Sentinel. Most people mix them up.
• An analytics rule runs KQL on your logs and produces an ALERT
• Related alerts get grouped (by entity, rule, time window) into an INCIDENT
• The incident is what the analyst actually owns: severity, status, owner, comments
Automation Rules act on incidents first: assign, tag, suppress known noise.
Playbooks (Logic Apps) do the heavy lifting: enrichment, notifications, containment.
Rule of thumb: triage logic in Automation Rules, integration logic in Playbooks. #MicrosoftSentinel
19
Most SOC teams don't need more alerts. They need fewer, better-enriched ones.
Before an analyst even opens an incident, automation should already have:
• IP and domain reputation
• Asset owner and criticality
• Past incidents for the same entity
That's where SOAR actually pays off. #SOC
1
40
Pulkit Rai S. retweeted
The IDE era is dead.
Introducing Reason, the AI Teammate for work and code
We made Reason manage customers, build itself and also make this video.
Pareto-frontier on Terminal Bench 4.0 and DeepSWE
Now you can try it for free -> reasonmachines.com
Pulkit Rai S. retweeted
Delete Chrome
Delete Granola
Delete Claude
Strawberry 1.0 is here, the AI browser that saves your team >100h/week
Pulkit Rai S. retweeted
LLMs operate on embeddings.
Humans speak in words.
What do tokens have to do with any of this!?
Well...




























