ip-api/proxy

WARNING: This is a pre-release version and is not suitable for production use. It is for evaluation purposes only. Please report all bugs by creating a issue, or by contacting us directly.

ip-api cache

A proxy that caches query results from the ip-api.com pro endpoint.

• better performance via multiple keep-alive connections to ip-api servers
• finds the best PoP based on latency
• retries failed requests
• advanced caching for each response field
• automatically batches requests to reduce network requests

Only /json and /batch are supported.

Getting Started

Install on Linux - Debian

• Create a user

adduser --system --disabled-password --disabled-login --home /opt/ip-api-proxy --group ip-api-proxy

• Install the latest version

If you do not have go installed, please see https://golang.org/doc/install.

GOBIN=/opt/ip-api-proxy/ go get -u github.com/ip-api/proxy/cmd/proxy

• Create a config file

/opt/ip-api-proxy/config

Example configuration:

IP_API_KEY=your_api_key
LOG_LEVEL=info

• Create a systemd file

/etc/systemd/system/ip-api-proxy.service

Suggested configuration:

[Unit]
Description=ip-api proxy server
Documentation=https://github.com/ip-api/proxy
After=network.target

[Service]
PermissionsStartOnly=true
LimitNOFILE=1048576
LimitNPROC=512
CapabilityBoundingSet=CAP_NET_BIND_SERVICE
AmbientCapabilities=CAP_NET_BIND_SERVICE
NoNewPrivileges=true
User=ip-api-proxy
WorkingDirectory=/opt/ip-api-proxy
EnvironmentFile=/opt/ip-api-proxy/config
ExecStart=/opt/ip-api-proxy/proxy
Restart=on-failure

[Install]
WantedBy=multi-user.target

• Enable and start the service

systemctl enable ip-api-proxy
systemctl start ip-api-proxy

Usage

Modify your applications to use http://127.0.0.1:8080 instead of http(s)://pro.ip-api.com.

Environment variables

Name	Type	Default	Description
IP_API_KEY	String	required	ip-api.com key
LISTEN	String	127.0.0.1:8080	ip:port to listen on
CACHE_TTL	Duration	24h	For how long to cache entries
CACHE_SIZE	Number	1073741824	In memory cache size
RETRIES	Number	4	How many times to retry backend requests
POPS_REFRESH	Duration	1h	How often to refresh the server locations
BATCH_DELAY	Duration	10ms	Max delay before sending a batch to the backend
LOG_OUTPUT	String	""	Set to "console" for console friendly output
LOG_LEVEL	String	""	Can be set to "info", "warn" or "error" to reduce log output
REVERSE_WORKERS	Number	10	How many workers to use for reverse lookups
REVERSE_PREFERGO	Bool	true	Prefer using Go's built-in DNS resolver

proxy's Issues

Kubernetes Deployment

My main use-case for this is to run this proxy on Kubernetes for our Logstash instances before they hit the ip-api api's. I am going to be creating a helm chart for the deployment of this with the Container that can now be built. Generally, within the helm community a helm chart lives in the repo of the code example: https://github.com/kubernetes/ingress-nginx/tree/master/charts/ingress-nginx
This would mean however that we would have to host the tar bundled charts someway so that people could use them.

Would this helm chart be something that you would like to have within this repo? I am more than happy to help maintain the chart.

TODO - Stats

Was reading the todo file about stats. Ideally, we would implement Prometheus metrics for this. OpenTelemetry is a good library to use todo this and implement other metrics exporters. Here is the go source code https://github.com/open-telemetry/opentelemetry-go

Container on DockerHub

Opening this issue to start the discussion about how we are wanting to have an official docker image on dockerhub (or some other container repo). Internally we use goreleaser which does all of our heavy lifting for releasing our go projects.