Flow Manager separates your flow json to multiple files
This assumes you have Node-RED already installed and working, if you need to install Node-RED see here
NOTE: This requires Node.js v8.11.1+
Feel free to donate to keep this project going!
https://www.paypal.com/donate/?business=6X6AADL3TVMZL&no_recurring=0¤cy_code=ILS
Install via Node-RED Manage Palette
node-red-contrib-flow-manager
Install via npm
$ cd ~/.node-red
$ npm install node-red-contrib-flow-manager
# Restart node-red
After installation of this module, during first boot only, A migration process will initiate.
During migration, your main flow json file will be split to multiple files which will store your node information.
It is advised to back up your main flow json, before running this module for the first time.
Node-RED startup process after migration:
All of your flow files are combined into a single JSON object, which is then loaded and served as your main flow object.
For that reason, it is advised to add your "fat" flow json file to .gitignore because from that moment, the flows are saved separately.
The nodes will be stored in the following subdirectories of your Node-RED path as such:
/flows/
flow name
/subflows/
subflow name
/config-nodes.json
(global config nodes will be stored here)
It's a good idea to add these paths to your version control system.
- Allows selecting which flows Node-RED will load, and which will be ignored and not loaded, not only in Node-RED's UI, also in it's NodeJS process.
- Unselected flows are NOT deleted, only "ignored" until you select them again using
Filter Flows
. - Filtering configuration is stored in
flow-manager-cfg.json
file under your Node-RED path. - if
flow-manager-cfg.json
file does not exist, or exists but malformed, or contains an empty JSON array, then all flows will be loaded and no filtering is done.
envnodes allows configuration of nodes using an external source and custom logic.
example:
create "envnodes/someName.jsonata" in your Node-RED project directory and put these contents:
(
$config := require(basePath & "/someConfig.json");
{
/* mqtt config node */
"21bcf36a.891e4d": {
"broker": $config.mqtt.broker,
"port": $config.mqtt.port
},
"name:MyInject": {
"topic": "niceTopic"
}
};
)
The result would be that your mqtt config node will use values from an external configuration file (which is useful in some cases), And a topic value will be inserted to an inject node matched with the given name.
Note basePath
is the path to your Node-RED directory.
When using Node-RED's "project mode", the value is the project folder path.
Supported envnode file ext: .jsonata .js .json
when using .js you can either module.exports = {...} the object directly, or use a function/async function (for example to read external file) and return an object.
js example:
const fs = require('fs-extra');
module.exports = async function() {
const configTopic = await fs.readJson('./config').topic;
return {
"name:MyInject": {
"topic": configTopic
},
"cff2a203.8158a": {
"someProperty": "myProperty"
}
}
}
Attempting to change any envnode controlled property via Node-RED UI/AdminAPI will be cancelled (with a warning popup) to keep original values defined in your envnodes configuration.
If you work in a server-client environment, where you work locally on your flows, and once in a while need to push your local changes to a remote process running Node-RED.
Do the following to enable Remote Deployment feature:
- Add remote definitions to
flow-manager-cfg.json
(you can add as many as you want, choose any name you wish for each remote)-
{ "filter": [], "fileFormat": "json", "remoteDeploy": { "remotes": [ { "name": "Production", "nrAddress": "http://yourAddress:1881" }, { "name": "Staging", "nrAddress": "http://yourOtherAddress:1881" } ] } }
-
- Restart Node-RED
- You should see the new "Remote Deploy" button on top of the Node-RED UI.
You can configure flow-manager to load/save flow files in YAML format instead oj JSON, by modifying file flow-manager-cfg.json
as such:
{
...
"fileFormat": "yaml"
...
}
The advantage is the code within function node becomes easier to read when inspecting the flow file itself.
See comparison below
- id: 493f0b3.b3b48f4
type: function
z: 70dd6be0.e35274
name: 'SomeFuncNode'
func: |-
const str = "hello";
console.log(str + ' world');
msg.payload = 'test';
return msg;
outputs: 1
noerr: 0
x: 580
'y': 40
wires:
- []
{
"id": "493f0b3.b3b48f4",
"type": "function",
"z": "70dd6be0.e35274",
"name": "SomeFuncNode",
"func": "const str = \"hello\";\nconsole.log(str + ' world');\nmsg.payload = 'test';\nreturn msg;",
"outputs": 1,
"noerr": 0,
"x": 580,
"y": 40,
"wires": [
[]
]
}
curl --request GET \
http://localhost:1880/flow-manager/flow-names
Response:
["Flow 1.json","Flow 2.json","Flow 3.json","Flow 4.json"]
curl --request GET \
http://localhost:1880/flow-manager/cfg
get
curl --request GET \
http://localhost:1880/flow-manager/filter-flows
change
curl --header "Content-Type: application/json" \
--request PUT \
--data '["Flow 1", "Flow 2"]' \
http://localhost:1880/flow-manager/filter-flows
Request URL template:
http://localhost:1880/flow-manager/flow-files/:type/:fileName?
type
can be flow
/subflow
/global
can pass plural too: flows
/subflows
Notice fileName
is optional if you request "global" flow, leave out the fileName
url part.
get
curl --request GET \
http://localhost:1880/flow-manager/flow-files/flow/MyFlow
delete
curl --request DELETE \
http://localhost:1880/flow-manager/flow-files/flow/MyFlow
add/update (mtime/atime query params are optional, in case you wish to set mtime "Modified Time" or atime "Access Time" after the file is written)
curl --header "Content-Type: application/json" \
--request POST \
--data '[{"id":"d4366369.5303d","type":"tab","label":"NewFlow","disabled":false,"info":""}]' \
http://localhost:1880/flow-manager/flow-files/flow/NewFlow
Example with mtime query param:
curl --header "Content-Type: application/json" \
--request POST \
--data '[{"id":"d4366369.5303d","type":"tab","label":"NewFlow","disabled":false,"info":""}]' \
'http://localhost:1880/flow-manager/flow-files/flows/NewFlow?mtime=8/16/2020,%206:12:17%20PM'
During runtime, you can request flow-manager to load any flow file placed in the flows directory,
by sending a POST request with the flows you wish to load (regardless of whether you configured any filter-flows or not)
Request to deploy all flows / redeploy current flows:
curl --header "Content-Type: application/json" \
--request POST \
--data '{"action":"loadAll"/"reloadOnly" }' \
http://localhost:1880/flow-manager/states
Request to add/remove/replace any flow:
curl --header "Content-Type: application/json" \
--request POST \
--data '{"action":"addOndemand"/"removeOndemand"/"replaceOndemand" "flows":["Flow 1","Flow 2"] }' \
http://localhost:1880/flow-manager/states
Retrieve state for specific type and flow.
http://localhost:1880/flow-manager/states/:type/:flowName
Retrieve state for all flows of a given type.
http://localhost:1880/flow-manager/states/:type
Retrieve state for all flows/subflows/global.
http://localhost:1880/flow-manager/states/
:type
can be subflow/flow/global (if global
, there's no need for :flowName
).
Example response for specific flow (http://localhost:1880/flow-manager/states/flow/Flow1
)
{
"deployed": true,
"hasUpdate": true,
"onDemand": false,
"rev": "477a9952b50e161afdf2e4bb6b84ee31",
"mtime": "2020-08-16T13:54:23.000Z",
"oldRev": "54d163c840657920aaac705dadecae2b", // If hasUpdate
"oldMtime": "2020-08-14T17:30:10.000Z" // If hasUpdate
}
deployed
whether that flow file is deployed.hasUpdate
whether that flow file was changed externally, and can be deployed again to update.onDemand
whether that flow was deployed by an OnDemand request, and is not part of filtered flows selection.rev/oldRev
revision (checksum) of file.mtime/oldMtime
Modified Time of file.
Request template:
http://localhost:1880/flow-manager/remotes/:remoteName/[append any rest endpoint here]
Example using flow-names
curl --request GET \
http://localhost:1880/flow-manager/remotes/Production/flow-names