Exporting metrics and check statuses to Prometheus

[thoo]

Is there a way to push metrics and statuses to Prometheus so that I can also visualize on Grafana?
Thanks.

[yisraelU]

you can push data via a webhook upon success or failure, you would need a proxy software to convert that to the format prometheus uses

[jhuckaby]

Although not documented, Cronicle does have an HTTP REST API for fetching metrics about all completed jobs in JSON format. I have no idea what data format Prometheus / Grafana use, but you could possibly use this as a starting point, and write a converter script or something. The URI endpoint is:

http://YOUR_SERVER:3012/api/app/get_history?api_key=YOUR_API_KEY&offset=0&limit=100

Make sure you create an API Key first, then use that in the URL above where it says YOUR_API_KEY. The offset and limit params are integers used for pagination through the timeline data (higher numbers go further back in time), so you can (and should) only fetch 100 or so records (completed jobs) at a time.

Example JSON response:

{
    "code": 0,
    "rows": [
        {
            "id": "jkd6rar2iog",
            "code": 0,
            "event": "ek0bnmr96hm",
            "category": "general",
            "plugin": "urlplug",
            "hostname": "mini.local",
            "time_start": 1595990100.618,
            "elapsed": 0.3429999351501465,
            "perf": {
                "scale": 1000,
                "perf": {
                    "total": 227.237,
                    "dns": 68.96,
                    "connect": 24.828,
                    "send": 52.286,
                    "wait": 73.618,
                    "receive": 1.224,
                    "decompress": 2.509
                },
                "counters": {
                    "bytes_sent": 154,
                    "bytes_received": 443
                }
            },
            "cpu": {},
            "mem": {},
            "log_file_size": 994,
            "action": "job_complete",
            "epoch": 1595990100,
            "event_title": "EMR Auto-Upload",
            "category_title": "General",
            "plugin_title": "HTTP Request"
        },
        {
            "id": "jkd6r4buzof",
            "code": 0,
            "event": "ejs9m57bw01",
            "category": "general",
            "plugin": "urlplug",
            "hostname": "mini.local",
            "time_start": 1595989800.971,
            "elapsed": 0.5130000114440918,
            "perf": {
                "scale": 1000,
                "perf": {
                    "total": 397.582,
                    "dns": 69.316,
                    "connect": 25.463,
                    "send": 53.506,
                    "wait": 241.533,
                    "receive": 1.244,
                    "decompress": 2.54
                },
                "counters": {
                    "bytes_sent": 172,
                    "bytes_received": 567
                }
            },
            "cpu": {},
            "mem": {},
            "log_file_size": 1150,
            "action": "job_complete",
            "epoch": 1595989801,
            "event_title": "AMR Auto-Upload",
            "category_title": "General",
            "plugin_title": "HTTP Request"
        },
        ...
]

Each item in the rows array is a completed job event, and contains lots of useful information you can graph. The code represents the exit status (0 means success, any other value means an error occurred). The perf object contains the performance metrics for the job, all in milliseconds. If the jobs lasted more than 10 seconds, you will also have the CPU and memory utilization data in the cpu and mem objects, respectively.

Good luck!

[thoo]

Thanks. That is what I am looking for.