Set up Gryphon for anything you run
Step-by-step guides for the services Gryphon has no check of its own for — Kafka, SQL Server, last night's backup — built from the checks it does have, with and without the agent.
15 of 18 guides
Show every guideDatabases
- Monitor MongoDB Check the port, then have the agent ask the server itself whether it answers, and whether its replica set has a primary. Agent optional TCP port · TCP port (agent) · Script (agent)
- Monitor Elasticsearch and OpenSearch Turn the cluster's green, yellow or red into Gryphon's healthy, warning or problem, with or without security switched on. Needs the agent HTTP (agent) · HTTPS (agent) · Script (agent)
Message brokers
- Monitor Kafka Know each broker is listening, and that your consumers are keeping up: consumer lag as a warning and a problem. Agent optional TCP port · TCP port (agent) · Script (agent) · Heartbeat
- Monitor RabbitMQ Confirm RabbitMQ itself answers on the AMQP port, with no credentials, and watch a queue that is backing up. Agent optional TCP send/expect · TCP send/expect (agent) · Script (agent)
Mail and network services
- Check any service with send/expect Knock on a port and listen for the answer only the real service gives. How to find that answer for anything that speaks over TCP, with tested settings for SSH, FTP, mail, Memcached, ZooKeeper, NATS, MQTT and more. Agent optional TCP send/expect · TCP send/expect (agent)
- Monitor a mail server SMTP, submission, IMAP and POP3, each checked for the greeting a real mail server gives, plus the certificates and the MX record. No agent needed TCP send/expect · SSL Certificate · DNS
Websites and apps
- Monitor your own app's health endpoint Point Gryphon at /healthz: from outside for a public app, through the agent for one on a private network, and require the answer to say it is healthy. Agent optional Monitor web page · HTTP (agent) · HTTPS (agent) · SSL Certificate · Script (agent)
- Monitor a WordPress site The page, its certificate and domain, the database behind it, and whether WordPress's scheduled tasks are actually running. Agent optional Monitor web page · SSL Certificate · Domain expiry · MariaDB/MySQL (agent) · Disk Space · Heartbeat
Alerts and integrations
- Send alerts to a Slack channel Make an incoming webhook in Slack, paste it into Gryphon, and every problem, warning, recovery and acknowledgement lands in the channel, on every plan. No agent needed Slack
- Send alerts to a Discord channel Create a webhook in your server's settings, paste it into Gryphon, and every problem, warning, recovery and acknowledgement is posted to the channel, without pinging anybody. No agent needed Discord
- Send alerts to a Microsoft Teams channel Add the Workflows template that posts webhook alerts to a channel, paste its URL into Gryphon, and alerts arrive as cards in the channel. The old connector URLs no longer work. No agent needed Microsoft Teams
- Receive alerts with a signed webhook Have Gryphon post every alert to your own HTTPS endpoint as JSON, signed so you can tell it came from Gryphon: the payload, the headers, and how to check the signature. No agent needed Webhook
Nothing here for what you run?
Every guide is made of the same six checks. Between them, they cover nearly anything that listens on a port, answers a request, can be asked by a script, or runs on a schedule.
Is anything listening?
TCP port, from outside or through the agent
Connects to a port and nothing more. Any service, any port: through the agent for an address only your own network can reach.
Is the right thing listening?
TCP send/expect, from outside or through the agent
Writes a line if you give it one, then looks for text in the answer, with TLS first for a service that needs it. Tells a mail server from whatever else is on port 25.
Does it answer properly?
Monitor web page, HTTP (agent), HTTPS (agent)
A request with the method you choose, the status codes you accept, and text the response must contain.
Anything you can script
Script (agent)
Exit 0, 1, 2 or 3 becomes healthy, warning, problem or unknown, and the first line of output is the message. Every Nagios plugin already works.
Did the job run?
Heartbeat, no agent
Your job calls a URL when it finishes, or another when it fails. Silence past its schedule is a problem, and so is a failure.
Did it write anything?
File freshness (agent)
How old a file is, or the newest file in a folder, and whether it is big enough. Catches the job that ran and wrote nothing.
The agent's checks need the agent on the machine, or on one in the same network. Install it in about a minute on Linux, Windows or a Mac, or read every check Gryphon has.
Fourteen days free. Then from $4.99 a month.
The agent, the dashboard, the apps and every check but the five for Kubernetes are in every plan. The plans differ in how much you watch, how often, from where, and how many people and status pages they include. Compare the plans. Cancel any time.
Already have an account? Sign in