The of-watchdog is a new version of the OpenFaaS watchdog which provides the original STDIO mode from the Classic Watchdog along with a new HTTP mode.
See also: Classic Watchdog
- Cleaner abstractions for maintenance
- Keep function process warm for lower latency / caching / persistent connections
- Explore streaming for large files (beyond disk/RAM capacity)
History/context: the original watchdog supported mode the Serializing fork mode only and Afterburn was available for testing via a pull request.
When the of-watchdog is complete this version will support four modes as listed below. We may consolidate or remove some of these modes before going to 1.0 so please consider modes 2-4 experimental.
The HTTP mode is set to become the default mode for future OpenFaaS templates.
The following templates have been available for testing:
| Template | HTTP framework | Repo |
|---|---|---|
| Node.js 8 | Express.js | https://github.com/openfaas-incubator/node8-express-template |
| Node.js 10 (LTS) | Express.js | https://github.com/openfaas-incubator/node10-express-template |
| Python 3 & 2.7 | Flask | https://github.com/openfaas-incubator/python-flask-template |
| Golang | Go HTTP (stdlib) | https://github.com/openfaas-incubator/golang-http-template |
| Golang | (http.HandlerFunc) | https://github.com/openfaas-incubator/golang-http-template |
| Ruby | Sinatra | https://github.com/openfaas-incubator/ruby-http |
| Java 8 | Sun HTTP / Maven | https://github.com/openfaas/templates/ |
Unofficial: .NET Core / C# and Kestrel
The HTTP mode is similar to AfterBurn.
A process is forked when the watchdog starts, we then forward any request incoming to the watchdog to a HTTP port within the container.
Pros:
-
Fastest option for high concurrency and throughput
-
More efficient concurrency and RAM usage vs. forking model
-
Database connections can be persisted for the lifetime of the container
-
Files or models can be fetched and stored in
/tmp/as a one-off initialization task and used for all requests after that -
Does not require new/custom client libraries like afterburn but makes use of a long-running daemon such as Express.js for Node or Flask for Python
Example usage for testing:
- Forward to an NGinx container:
$ go build ; mode=http port=8081 fprocess="docker run -p 80:80 --name nginx -t nginx" upstream_url=http://127.0.0.1:80 ./of-watchdog
- Forward to a Node.js / Express.js hello-world app:
$ go build ; mode=http port=8081 fprocess="node expressjs-hello-world.js" upstream_url=http://127.0.0.1:3000 ./of-watchdog
Cons:
-
One more HTTP hop in the chain between the client and the function
-
Daemons such as express/flask/sinatra can be unpredictable when used in this way so many need additional configuration
-
Additional memory may be occupied between invocations vs. forking model
This mode is designed to replicate the behaviour of the original watchdog for backwards compatibility.
Forks one process per request. Multi-threaded. Ideal for retro-fitting a CGI application handler i.e. for Flask.
Limited to processing files sized as per available memory.
Reads entire request into memory from the HTTP request. At this point we serialize or modify if required. That is then written into the stdin pipe.
-
Stdout pipe is read into memory and then serialized or modified if necessary before being written back to the HTTP response.
-
A static Content-type can be set ahead of time.
-
HTTP headers can be set even after executing the function (not implemented).
-
Exec timeout: supported.
Forks a process per request and can deal with a request body larger than memory capacity - i.e. 512mb VM can process multiple GB of video.
HTTP headers cannot be sent after function starts executing due to input/output being hooked-up directly to response for streaming efficiencies. Response code is always 200 unless there is an issue forking the process. An error mid-flight will have to be picked up on the client. Multi-threaded.
-
Input is sent back to client as soon as it's printed to stdout by the executing process.
-
A static Content-type can be set ahead of time.
-
Exec timeout: supported.
Afterburn should be considered for deprecation in favour of the HTTP mode.
Several sample templates are available under the OpenFaaS incubator organisation.
https://github.com/openfaas/nodejs-afterburn
https://github.com/openfaas/python-afterburn
https://github.com/openfaas/java-afterburn
Uses a single process for all requests, if that request dies the container dies.
Vastly accelerated processing speed but requires a client library for each language - HTTP over stdin/stdout. Single-threaded with a mutex.
-
Limited to processing files sized as per available memory.
-
HTTP headers can be set even after executing the function.
-
A dynamic Content-type can be set from the client library.
-
Exec timeout: not supported.
Environmental variables:
Note: timeouts should be specified as Golang durations i.e.
1mor20s.
| Option | Implemented | Usage |
|---|---|---|
function_process |
Yes | Process to execute a server in http mode or to be executed for each request in the other modes. For non http mode the process must accept input via STDIN and print output via STDOUT. Alias: fprocess |
read_timeout |
Yes | HTTP timeout for reading the payload from the client caller (in seconds) |
write_timeout |
Yes | HTTP timeout for writing a response body from your function (in seconds) |
exec_timeout |
Yes | Exec timeout for process exec'd for each incoming request (in seconds). Disabled if set to 0. |
port |
Yes | Specify an alternative TCP port for testing. Default: 8080 |
write_debug |
No | Write all output, error messages, and additional information to the logs. Default is false. |
content_type |
Yes | Force a specific Content-Type response for all responses - only in forking/serializing modes. |
suppress_lock |
Yes | When set to false the watchdog will attempt to write a lockfile to /tmp/ for healthchecks. Default false |
upstream_url |
Yes | http mode only - where to forward requests i.e. 127.0.0.1:5000 |
buffer_http |
Yes | http mode only - buffers request body to memory before fowarding. Use if your upstream HTTP server does not accept Transfer-Encoding: chunked Default: false |
max_inflight |
Yes | Limit the maximum number of requests in flight |
Note: the .lock file is implemented for health-checking, but cannot be disabled yet. You must create this file in /tmp/.