Add job_event and job_event_ack to the gateway protocol

A gateway keeps what happened to a job it was handed (started, finished,
failed) and re-sends it on every connection until the cloud acknowledges
it, so an outcome reached while the cloud was unreachable still arrives.

schemas/continuum/v1/edge/client_message.schema.json

@@ -195,6 +195,20 @@
       },
       "required": ["type", "request_id"]
     },
+    {
+      "type": "object",
+      "title": "JobEvent",
+      "description": "What happened to a job this gateway was handed, kept in the gateway's own database and re-sent on every connection until the cloud answers with a JobEventAck - so an outcome reached while the cloud was unreachable (or the gateway restarted) still arrives. The cloud applies it only to a job that is still in flight, so a job it has since cancelled or reassigned is left alone, and applying the same event twice changes nothing.",
+      "properties": {
+        "type": { "const": "job_event" },
+        "event_id": { "type": "integer", "description": "The gateway's own sequence number for this event; echoed in the JobEventAck." },
+        "job_id": { "type": "string", "format": "uuid" },
+        "event": { "enum": ["started", "finished", "failed"] },
+        "detail": { "type": ["string", "null"], "description": "For failed: the reason, in words for an operator." },
+        "at_ms": { "type": "integer", "description": "When the gateway observed it, milliseconds since the Unix epoch." }
+      },
+      "required": ["type", "event_id", "job_id", "event", "at_ms"]
+    },
     {
       "type": "object",
       "title": "DiscoveredPrinters",