/nucypher-monitoring

NuCypher Worker Node Monitoring Suite

Primary LanguageShellApache License 2.0Apache-2.0

NuCypher Worker Node Monitoring Suite

This is an opinionated, inexpensive and simple to set up suite of NuCypher node monitoring and alerting services using Prometheus and Grafana. It's boiled down to be of use to the less techincal node operators out there: in a professional setting we expect the users to have more sophisticated setups.

Disclaimer

This guide was wrote for persons who complete NuCypher installation guide using SystemD service installation and set --prometheus --metrics-port 9101 parameters
The guide was tested on Ubuntu 18.04 and should be applicable to all modern debian-based distributions.

Contribution policy

Feel free to open issues or pull requests, but we won't accept PRs that make monitoring setup or maintenance significantly more difficult or expensive.

Pre-Installation

  1. Clone the repository

    git clone https://github.com/p2p-org/nucypher-monitoring.git

Installation

  1. Install docker and docker-compose if you didn't it before. It needs root privileges to run (so, run it using sudo or under root user).

    curl -fsSL https://get.docker.com -o- | sh

curl -fsSL https://github.com/docker/compose/releases/download/1.25.4/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose && \
chmod +x /usr/local/bin/docker-compose && \
ln -s /usr/local/bin/docker-compose /usr/bin/docker-compose && \
curl -fsSL https://raw.githubusercontent.com/docker/compose/1.25.4/contrib/completion/bash/docker-compose -o /etc/bash_completion.d/docker-compose

  2. Install node-exporter, set listening address to 127.0.0.1 and enable expose systemd metrics. You can use bash-script install-prometheus-node-exporter.sh to do all that automaticalle. It needs root privileges to run.

    ./install-prometheus-node-exporter.sh

Setting up environment

NB: this setup exposes grafana publicly over HTTP. For a more secure setup consider getting a domain and setting up SSL or using ngrok.

You need to setup environment variables.

  1. Open file grafana.env in any text editor;
  2. Replace 0.0.0.0 in GF_SERVER_ROOT_URL with your server's external address or DNS name for correct url in alert notifications;
  3. If you want to receive email notifications then you need to configure SMTP settings:
    • Replace host:port in GF_SMTP_HOST with your SMTP server DNS name or IP and port;
    • Set GF_SMTP_USER,GF_SMTP_FROM_ADDRESS and GF_SMTP_PASSWORD to Email/Password from which the notifications will be sent.
  4. Save changes.

For example, for the address admin@gmail.com:

GF_SMTP_HOST=smtp.gmail.com:465
GF_SMTP_USER=admin@gmail.com
GF_SMTP_PASSWORD=examplePassword
GF_SMTP_FROM_ADDRESS=admin@gmail.com

Do not use your main e-mail account for this, register a throwaway.

Launch

  1. Run docker-compose.

    docker-compose up -d

Now you can see your dashboard at http://your.server.ip.address.or.dns.name:3000/. In a fresh grafana isntall, you will be prompted for username/password. Use username admin with password admin. You will be prompted to change your password.

Configure alert channels

Telegram bot alerting

  1. Get telegram bot token by using appropriate telegram bot guide, also get your chat id (e.g. via @get_id_bot, see additional guides in grafana community);

  2. Go to grafana Alerting -> Notification channels;

  3. Click on New channel button;

  4. Select Type -> Telegram;

  5. Fill Name, paste your bot token to BOT API Token field and your chat id to Chat ID field. Click on Send Test to check the sending of alerts and if it's OK click on Save button;

  6. Additionally you can set another options for including image into messages, send reminders, etc., like on screenshot below.

OpsGenie

OpsGenie allows for mobile network based alerts like SMS or phone calls. You can use an other similar service for that.

  1. Get API key as described in OpsGenie documentation;
  2. Repeat items 2, 3 on the list above;
  3. Select Type -> OpsGenie;
  4. Fill Name, paste your API key to API key. Click on Send Test to check the sending of alerts and if it's OK click on Save button;

Email

NB: for e-mail alerting to work you need to set up email-related environment variables, as described in Setting up environment

  1. Go to grafana Alerting -> Notification channels;

  2. Click on New channel button;

  3. Select Type -> Email;

  4. Fill Name and fill in the Addresses field with the Emails to which notifications will be received.

Configure external HTTP monitoring

We highly recommend to configure external HTTP monitoring for grafana dashboard and ursula's status page. That way you'll recieve an alert even if your server, ursula node or grafana service goes down. We describe using Uptime Robot because it has a pretty good entry level free plan at the time of the writing, but there are dozens of services like that that could be used for the purpose.

UptimeRobot.com

Configure alert contacts

Email

  1. Go to My settings (1) and click on Create Alert Contact (2).

  2. Fill all fields like on screenshot.

  3. Click on Create Alert contact (4).

Telegram

  1. Go to My settings and click on Create Alert Contact.

  2. Fill all fields like on screenshot.

  3. Click on Create Alert contact.

  4. Click on custom link (see screenshot). It will open Telegram chat with UptimeRobot bot.

  5. Type command /start.

Create monitor

Create one monitor for Ursula's status page URL https://YOUR_SERVER_EXTERNAL_IP:9151/status/ and another one for grafana dashboard URL (e.g. http://YOUR_SERVER_EXTERNAL_IP:3000/)

  1. Go to dashboard, click on Add New Monitor.

  2. Select Monitor type -> HTTP(s) (1).

  3. Fill Friendly name (2), URL (or IP) (3), and select needed Alert contacts (4)(5).

  4. Click on Create Monitor (6).