@daemonitor/plugins 1.1.20 → 1.1.23

package/dist/plugins/monitoring/DockerPlugin.js

@@ -128,6 +128,38 @@ async function collectContainers(bin) {
     catch (err) {
         console.error("docker: inspect collection failed:", err.message);
     }
+    // 4. For UNHEALTHY containers only, read the healthcheck's own last output.
+    //
+    // "1/13 containers unhealthy" says a problem exists and nothing about what it
+    // is. Docker already stores the failing command's output in
+    // .State.Health.Log[] — that is the actual error text, and without it an alert
+    // sends someone to ssh into the box to run the same healthcheck by hand.
+    //
+    // Read one container at a time and only for the unhealthy ones. The batched
+    // inspect above avoids .State.Health because the template errors for
+    // containers that have no healthcheck; a container reporting "(unhealthy)"
+    // necessarily has one, so this is safe, and it is a handful of calls at most
+    // precisely when something is wrong.
+    const unhealthy = containers.filter((c) => c.health === "unhealthy").slice(0, 5);
+    for (const c of unhealthy) {
+        try {
+            const { stdout } = await execAsync(`${bin} inspect --format '{{json .State.Health}}' ${c.id}`, EXEC_OPTS);
+            const h = JSON.parse(stdout.trim() || "null");
+            if (!h)
+                continue;
+            c.healthFailingStreak = Number(h.FailingStreak) || 0;
+            const last = Array.isArray(h.Log) && h.Log.length ? h.Log[h.Log.length - 1] : null;
+            if (last?.Output) {
+                // Healthcheck output is often multi-line and ends in a newline; collapse
+                // it so it fits an alert line, and keep the head where the error is.
+                c.healthOutput = String(last.Output).replace(/\s+/g, " ").trim().slice(0, 300);
+            }
+        }
+        catch (err) {
+            // Best-effort: an alert naming the container is still better than none.
+            console.error(`docker: health log read failed for ${c.name}:`, err.message);
+        }
+    }
     return containers;
 }
 export function createDockerPlugin() {

package/dist/plugins/monitoring/SnmpPlugin.js

@@ -92,7 +92,7 @@ function decodeSnmp(s) {
 }
 async function snmpGet(t, oid, timeout = 8000) {
     try {
-        const { stdout } = await pexec("snmpget", ["-v", t.version || "2c", "-c", t.community || "public", "-Ovq", "-t", "2", "-r", "1", t.ip, oid], { timeout });
+        const { stdout } = await pexec("snmpget", ["-v", t.version || "2c", "-c", t.community || "public", "-Oevq", "-t", "2", "-r", "1", t.ip, oid], { timeout });
         const raw = stdout.trim().replace(/^"|"$/g, "");
         if (!raw || /No Such|no response|Timeout/i.test(raw))
             return null;
@@ -104,7 +104,7 @@ async function snmpGet(t, oid, timeout = 8000) {
 }
 async function snmpWalk(t, oid, timeout = 10000) {
     try {
-        const { stdout } = await pexec("snmpwalk", ["-v", t.version || "2c", "-c", t.community || "public", "-Oqv", "-t", "2", "-r", "1", t.ip, oid], { timeout });
+        const { stdout } = await pexec("snmpwalk", ["-v", t.version || "2c", "-c", t.community || "public", "-Oeqv", "-t", "2", "-r", "1", t.ip, oid], { timeout });
         return stdout.split("\n").map((l) => decodeSnmp(l.trim().replace(/^"|"$/g, ""))).filter(Boolean);
     }
     catch {
@@ -184,14 +184,22 @@ async function collectDevice(t) {
     };
     const operByIdx = byIndex(oper);
     const adminByIdx = byIndex(admin);
+    // Values arrive numeric (-Oe), but tolerate an enum rendering like "up(1)"
+    // from any agent/MIB combination that slips one through.
+    const enumNum = (v) => {
+        if (v == null)
+            return undefined;
+        const m = /\((\d+)\)\s*$/.exec(v) || /^\s*(-?\d+)\s*$/.exec(v);
+        return m ? Number(m[1]) : undefined;
+    };
     // Only count ADMIN-enabled ports; an enabled port that's operationally down is
     // a real link-down (ignore the many disabled ports on a big switch).
     let portsTotal = 0;
     let portsUp = 0;
     for (const idx of Object.keys(adminByIdx)) {
-        if (Number(adminByIdx[idx]) === 1) {
+        if (enumNum(adminByIdx[idx]) === 1) {
             portsTotal++;
-            if (Number(operByIdx[idx]) === 1)
+            if (enumNum(operByIdx[idx]) === 1)
                 portsUp++;
         }
     }
@@ -267,8 +275,8 @@ async function collectDevice(t) {
         }
         // ifOperStatus 1=up, ifAdminStatus 1=enabled. A link that's admin-enabled but
         // operationally down is a real fault; an admin-disabled port is just off.
-        const adminUp = adminByIdx[idx] != null ? Number(adminByIdx[idx]) === 1 : undefined;
-        const linkUp = operByIdx[idx] != null ? Number(operByIdx[idx]) === 1 : undefined;
+        const adminUp = adminByIdx[idx] != null ? enumNum(adminByIdx[idx]) === 1 : undefined;
+        const linkUp = operByIdx[idx] != null ? enumNum(operByIdx[idx]) === 1 : undefined;
         interfaces.push({ name, inBps, outBps, up: linkUp, adminUp });
     }
     netPrevIf[t.ip] = nextIf;
@@ -306,7 +314,7 @@ async function collectDevice(t) {
 // hrStorage table entries for the data volumes.
 async function snmpWalkIndexed(t, oid, timeout = 10000) {
     try {
-        const { stdout } = await pexec("snmpwalk", ["-v", t.version || "2c", "-c", t.community || "public", "-On", "-t", "2", "-r", "1", t.ip, oid], { timeout });
+        const { stdout } = await pexec("snmpwalk", ["-v", t.version || "2c", "-c", t.community || "public", "-Oen", "-t", "2", "-r", "1", t.ip, oid], { timeout });
         return stdout.split("\n").map((l) => {
             const m = /^(\S+)\s*=\s*[^:]+:\s*(.*)$/.exec(l.trim());
             if (!m)

package/package.json

@@ -1,15 +1,10 @@
 {
   "name": "@daemonitor/plugins",
-  "version": "1.1.20",
+  "version": "1.1.23",
   "description": "daemonitor plugins",
   "types": "types/index.d.ts",
   "main": "dist/index.js",
   "type": "module",
-  "scripts": {
-    "build": "tsc",
-    "test": "echo \"Error: no test specified\" && exit 1",
-    "prepublishOnly": "npm run build"
-  },
   "files": [
     "dist",
     "types"
@@ -53,5 +48,9 @@
   "bugs": {
     "url": "https://github.com/daemonitor/plugins/issues"
   },
-  "homepage": "https://github.com/daemonitor/plugins#readme"
-}
+  "homepage": "https://github.com/daemonitor/plugins#readme",
+  "scripts": {
+    "build": "tsc",
+    "test": "echo \"Error: no test specified\" && exit 1"
+  }
+}