Memo

inspired by https://github.com/Dieterbe/anthracite/ but now as part of your chatops, chatdev, chatmarketing, chatWhatever workflow!

Comes with 2 programs:

• memo-cli: submit grafana annotations from the cli
• memod: slack bot, so you can can submit annotations from slack

Huh?

Turn a slack message like this ... ... into an annotation like this: Luckily somebody shared this memo on slack, otherwise somebody might freak out if they see this chart!

memo-cli

Usage of ./memo-cli:
  -config string
    	config file location (default "~/.memo.toml")
  -msg string
    	message to submit
  -tags value
    	One or more comma-separated tags to submit, in addition to 'memo', 'user:<unix-username>' and 'host:<hostname>'
  -ts int
    	unix timestamp. always defaults to 'now' (default 1557953985)

memod

Connects to slack and listens for "memo" messages which - if correctly formatted - will result in an annotation on the configured Grafana server

Message format

memo [timespec] <msg> [tags]

[foo] denotes that foo is optional.

timespec

defaults to 25, so by default it assumes your message is about 25 seconds after the actual event happened.

It can have the following formats:

• <duration> like 0 (seconds), 10 (seconds), 30s, 1min20s, 2h, etc. see https://github.com/raintank/dur denotes how long ago the event took place
• <RFC3339 spec> like 2013-06-05T14:10:43Z

msg

free-form text message, but if the first word looks like a timespec it will be interpreted as such. Any words at the end with : in them will be interpreted as tags.

tags

default tags included:

• memo
• chan:slack channel (if not a PM)
• author:slack username

you can extend these. any words at the end of the command that have : will be used as key-value tags. But you cannot override any of the default tags

Installation

Configure slack (only for memod)

You set up the slack bot as a bot integration. Essentially in your slack workspace configuration, you create a bot with the name "memobot" and a token, and specify which channels to join by default. Put the token in the memod config file (see below) and it should just work. In the future we may look into creating a "real slack app". Note that after the bot joins, you can still invite it into - or remove from - any other channel.

1. Create an app token
2. Enable socket mode on your application
3. Enable event subscriptions
4. Subscribe to message.channels and message.im
5. Create a bot token (in OAuth and Permissions)

Scopes required for bot token:

• channels:history
• channels:read
• chat:write
• im:history
• im:read
• users:read

Install the program

Currently we don't publish distribution packages, docker images etc. So for now, you need to build the binary/binaries from source

First, install golang Then, run any of these commands to download the source code and build the binaries:

go get github.com/grafana/memo/cmd/memod    # only memod, the slack bot
go get github.com/grafana/memo/cmd/memo-cli # only memo-cli, the command line tool
go get github.com/grafana/memo/cmd/...      # both

You will then have the binaries in $HOME/bin or in $GOPATH/bin if you have a custom GOPATH set.

config file for memo-cli

Put this file in ~/.memo.toml

[grafana]
api_key = "<grafana api key, editor role>"
api_url = "https://<grafana host>/api/"

config file for memod

Put a config file like below in /etc/memo.toml.

# one of trace debug info warn error fatal panic
log_level = "info"

[slack]
enabled = true
bot_token = "<slack bot token>"
app_token = "<slack app token>"

[discord]
enabled = true
bot_token = "<discord bot token>"

[grafana]
api_key = "<grafana api key, editor role>"
api_url = "http://localhost/api/"

auto-starting memod

If you use upstart, you need to create an init file and put it in /etc/init/memo.conf For your convenience you can use our example upstart config file In this case also copy the binary to /usr/bin/memod.

Set up the Grafana integration

You need to create a new annotation query on your applicable dashboards. Make sure to set it to the Grafana datasource and use filtering by tag, you can use tags like memo and chan:<chan-name> or any other tags of your choosing.

Docker

A docker image is compiled for convenience:

# memod
docker run -v "${PWD}/config.toml:/etc/memo.toml" ghcr.io/grafana/memo:master

# memo-cli
docker run -v "${PWD}/config.toml:/etc/memo.toml" ghcr.io/grafana/memo:master memo-cli -config /etc/memo.toml -msg "test"