pathivu / pathivu

An efficient log ingestion and log aggregation system https://pathivu.io/

Pathivu: Logs you can search

Pathivu is a powerful and lightweight log ingestion and aggregation system. It offers cost-efficiency and high throughput without trading away log indexing. It is perfect for cloud native workloads.

Index

• Highlights
• Architecture
• Pathivu Server
• Pathivu Web
• Katchi
• Usage
• Documentation
• Website

Highlights

• Fast Ingestion
• Beautiful dashboard*
• Log tailing
• Log indexing
• Cost-Efficient
• Intuitive query language
• Multi-threaded log ingestion
• Structured logging replayer
• Log retention

Architecture

A fluentd service running on a Kubernetes node can be used to ship logs to the Pathivu server. The server then exposes two types of interfaces, namely web and CLI. The following are the components of Pathivu:

• Pathivu server: A high performant log ingestion server where you can ingest logs at terabyte scale and query logs in efficient way, with both HTTP(s) as well as gRPC endpoints available. Read more about it here

• Katchi CLI: Katchi is the command line interface for Pathivu. Read more about it here

• Pathivu Web: It is a web based user interface for interacting with Pathivu. It makes the querying experience simplistic and intuitive. Read more about it here

Pathivu Server

Pathivu server is a high performant log ingestion server where you can ingest logs at terabyte scale and query logs in efficient way, with both HTTP(s) as well as gRPC endpoints available. It offers a gRPC service for fast log ingestion and an HTTP(s) backend for log querying and aggregation. By default, log ingestion runs on gRPC port 6180 and querying on HTTP(s) port 5180.

Pathivu server has a fluentd connector that can be used to ship logs to pathivu from a Kubernetes node. For more information on architecture, see this.

Pathivu Web

Pathivu Web is a web based user interface for interacting with Pathivu. It makes the querying experience simplistic and intuitive. You can use Pathivu's simplistic query language to interact with the server right from your browser.

Katchi

Katchi CLI can be used to interact with the pathivu server instance. It has the following functionalities:

• Log service discovery
• Querying
• Tailing
• Log output

For example, the distinct command also provides a feature to count the number of distinct logs matched. It is a very powerful query which can handle data at a terabyte scale. For the following JSON, the distinct(level)command will give you a list of all distinct levels in the logs.

{
  "data": [
    {
      "ts": 3,
      "entry": {
        "details": {
          "error_code": "500",
          "message": "Invalid URI"
        },
        "level": "warn",
        "from": "backend"
      },
      "source": "demo"
    },
    {
      "ts": 2,
      "entry": {
        "details": {
          "error_code": "500",
          "message": "Error connecting to database"
        },
        "level": "fatal",
        "from": "app"
      },
      "source": "demo"
    }
  ]
}

So the output will look something like this:

{
  "data": [
    "fatal",
    "warn"
  ]
}

Katchi connects to your pathivu server instance for live log tailing as well as viewing a log snapshot. It can be triggered in the following way:

$ katchi logs --host=http://localhost:5180
APP: kube-addon-manager-minikube, ts: 2019-11-18 00:06:58 +0530 IST, line: INFO: == Kubernetes addon reconcile completed at 2019-11-17T18:36:58+00:00 ==
 
APP: kube-addon-manager-minikube, ts: 2019-11-18 00:07:02 +0530 IST, line: INFO: Leader election disabled.
 
APP: kube-addon-manager-minikube, ts: 2019-11-18 00:07:03 +0530 IST, line: INFO: == Kubernetes addon ensure completed at 2019-11-17T18:37:03+00:00 ==
 
APP: kube-addon-manager-minikube, ts: 2019-11-18 00:07:03 +0530 IST, line: INFO: == Reconciling with deprecated label ==
 
APP: kube-addon-manager-minikube, ts: 2019-11-18 00:07:03 +0530 IST, line: error: no objects passed to apply

Learn more about Katchi here

Usage

Pathivu can be deployed to your own kubernetes cluster, just follow the steps mentioned below.

# Create a namespace
kubectl create -f https://raw.githubusercontent.com/pathivu/pathivu/master/kubernetes/debug/namespace.yaml


# Create pathivu service
kubectl create -f https://raw.githubusercontent.com/pathivu/pathivu/master/kubernetes/debug/service.yaml

# Create pathivu deployment
kubectl create -f https://raw.githubusercontent.com/pathivu/pathivu/master/kubernetes/debug/deployment.yaml

Pathivu has an internal fluentd connector that can be used for log ingestion. The following command initialized the connector and starts shipping your service logs to pathivu.

kubectl create -f https://raw.githubusercontent.com/pathivu/pathivu/master/kubernetes/debug/fluentd.yaml

pathivu's Issues

crash replayer

landing page and doc site.

trim down dependecies

structured logging

do grace ful shut down

backward iteration

build dashboard

Integrate CI

make a tool to insert log line from file for testing

Priority: Fix katchi for tailing and structed logs

add open tracing

Add support for distributed pathivu

Features for 0.1

• structured log baseline(only ingestion)
• backward iteration
• crash replayer
• tailing
• katchi live tailing