sap-health-triagelisted

# SAP Health Triage The "is it up and healthy — and if not, why?" skill. Mostly **read-only** — safe to run — but confirm the **SID** before pointing at a system, and note that several SAPControl methods are **protected** and may need to be permitted/authenticated (§7). ## Guardrail note Health checks don't change state, so the full stop/delete guardrails don't apply — but still: **Identify** the SID/host/instance before running, and treat any command that *writes* (none here) with the repo guardrails. Reading another system's syslog/traces is still accessing that system — confirm you have the right SID. --- ## 0. When the system is jammed — the out-of-band path **Read this first if SAP GUI will not log on, SM50/SM21 are unreachable, or RFC calls hang.** > ## Why `sapcontrol` still answers when the system does not > > **`sapstartsrv` is a separate operating-system process.** It is not an ABAP work process, it does > not queue behind the dispatcher, and it does not need a free dialog work process to serve a > request. It reads the instance's **shared memory** and files directly and answers over its own > SOAP/HTTP port (`5<nr>13` / `5<nr>14`). > > **That is the whole reason this path exists:** when every work process is occupied and the > dispatcher queue is full, the ABAP stack cannot answer *anything* — but `sapcontrol` can still tell > you the work-process table, the syslog and the queue depth. It is your **out-of-band** channel, the > equivalent of a lights-out cons