Failure to start RedisClientSupervisor child process - Elixir + Phoenix - redis

Question

I'm running an application with Elixir 1.7.2 and Phoenix but it randomly crashed with errors in one of my child processes.
I'm using a module called redix - {:redix, ">= 0.6.1"} which is handling my connection to Redis.
It crashed with the following error:
2019-05-16T07:29:08.929601+00:00 app[web.1]: 07:29:08.929 [info] Running ReviewedoWeb.Endpoint with Cowboy using http://:::46402
2019-05-16T07:29:09.191365+00:00 app[web.1]: 07:29:09.191 [info] Application reviewedo exited: Reviewedo.Application.start(:normal, []) returned an error: shutdown: failed to start child: Reviewedo.RedisClientSupervisor
2019-05-16T07:29:09.191385+00:00 app[web.1]: ** (EXIT) shutdown: failed to start child: Reviewedo.Accounts.Cache
2019-05-16T07:29:09.191387+00:00 app[web.1]: ** (EXIT) an exception was raised:
2019-05-16T07:29:09.191393+00:00 app[web.1]: ** (ArgumentError) the user in the Redis URI is ignored and should not be present, got: "h"
2019-05-16T07:29:09.191394+00:00 app[web.1]: (redix) lib/redix/uri.ex:30: Redix.URI.password/1
2019-05-16T07:29:09.191396+00:00 app[web.1]: (redix) lib/redix/uri.ex:15: Redix.URI.opts_from_uri/1
2019-05-16T07:29:09.191397+00:00 app[web.1]: (redix) lib/redix.ex:290: Redix.start_link/2
2019-05-16T07:29:09.191398+00:00 app[web.1]: (stdlib) supervisor.erl:365: :supervisor.do_start_child/2
2019-05-16T07:29:09.191399+00:00 app[web.1]: (stdlib) supervisor.erl:348: :supervisor.start_children/3
2019-05-16T07:29:09.191400+00:00 app[web.1]: (stdlib) supervisor.erl:314: :supervisor.init_children/2
2019-05-16T07:29:09.191402+00:00 app[web.1]: (stdlib) gen_server.erl:365: :gen_server.init_it/2
2019-05-16T07:29:09.191403+00:00 app[web.1]: (stdlib) gen_server.erl:333: :gen_server.init_it/6
2019-05-16T07:29:10.716373+00:00 app[web.1]: {"Kernel pid terminated",application_controller,"{application_start_failure,reviewedo,{{shutdown,{failed_to_start_child,'Elixir.Reviewedo.RedisClientSupervisor',{shutdown,{failed_to_start_child,'Elixir.Reviewedo.Accounts.Cache',{'EXIT',{#{'__exception__' => true,'__struct__' => 'Elixir.ArgumentError',message => <<\"the user in the Redis URI is ignored and should not be present, got: \\"h\\"\">>},[{'Elixir.Redix.URI',password,1,[{file,\"lib/redix/uri.ex\"},{line,30}]},{'Elixir.Redix.URI',opts_from_uri,1,[{file,\"lib/redix/uri.ex\"},{line,15}]},{'Elixir.Redix',start_link,2,[{file,\"lib/redix.ex\"},{line,290}]},{supervisor,do_start_child,2,[{file,\"supervisor.erl\"},{line,365}]},{supervisor,start_children,3,[{file,\"supervisor.erl\"},{line,348}]},{supervisor,init_children,2,[{file,\"supervisor.erl\"},{line,314}]},{gen_server,init_it,2,[{file,\"gen_server.erl\"},{line,365}]},{gen_server,init_it,6,[{file,\"gen_server.erl\"},{line,333}]}]}}}}}},{'Elixir.Reviewedo.Application',start,[normal,[]]}}}"}
2019-05-16T07:29:10.717128+00:00 app[web.1]: Kernel pid terminated (application_controller) ({application_start_failure,reviewedo,{{shutdown,{failed_to_start_child,'Elixir.Reviewedo.RedisClientSupervisor',{shutdown,{failed_to_start_child,'Elixir
2019-05-16T07:29:10.719206+00:00 app[web.1]:
2019-05-16T07:29:13.269465+00:00 heroku[web.1]: State changed from up to crashed
2019-05-16T07:29:13.189754+00:00 app[web.1]: Crash dump is being written to: erl_crash.dump...done
2019-05-16T07:29:13.256383+00:00 heroku[web.1]: Process exited with status 1
I hadn't made any source code changes before the crash and so I'm not sure what the cause is.
The error I'm seeing is this one:
https://github.com/whatyouhide/redix/blob/master/lib/redix/uri.ex#L31
It looks like it's trying to split my redis_url to get the user and password but it's coming up with h for the user. The redis_url that was generated for me looks similar to this:
redis://h:pf851f7d83mf93nrand0mha5hsh8kgl3h4334k73d82580bba59d4c237f33f24752c76c#ec2-45-81-27-450.region.compute.amazonaws.com:18324
Here are the modules that it is erroring on:
defmodule Reviewedo.RedisClientSupervisor do
#moduledoc """
Initiates Redis
"""
use Supervisor
alias Reviewedo.Accounts.Cache
def start_link do
Supervisor.start_link(__MODULE__, [])
end
def init(_) do
redis_url = System.get_env("REDIS_URL") || "redis://127.0.0.1:6379"
children = [
worker(Cache, [redis_url, :redix], [])
]
supervise(children, strategy: :one_for_one)
end
end
defmodule Reviewedo.Accounts.Cache do
#moduledoc """
Provides cache functionality
"""
def start_link(connection, name) do
Redix.start_link(connection, name: name)
end
def query(param), do: Redix.command(:redix, param)
def get(key), do: query(["get", key])
def set(key, value), do: query(["set", key, value])
def expire(key, seconds), do: query(["expire", key, seconds])
def del(key), do: query(["del", key])
def all_keys!() do
case query(["keys", "*"]) do
{:ok, keys} -> keys
_ -> raise Reviewedo.RedisConnectionError, query: "all_keys!/1"
end
end
def flushdb do
query(["flushdb"])
:ok
end
end
defmodule Reviewedo.RedisConnectionError do
#moduledoc """
Raised when a Cache query returns an error
This exception is raised by `Reviewedo.Accounts.Cache` and `Reviewedo.Accounts.CacheAPI` which:
"""
defexception [:message]
def exception(query: query) do
msg = "Redis connection error, calling query: `#{query}`"
%Reviewedo.RedisConnectionError{message: msg}
end
end
It looks as though it's failing to get my opts from the uri but I'm not quite sure why. The application is running fine locally so it must be something to do with the configuration on Heroku.
Has anyone experienced anything similar to this before or know what the problem might be?
Thanks in advance!