Pedix Logging Overview
About Predix Logging
Learn about the Predix Logging service
Use the Predix Logging service to collate and view your log files. Predix Logging is the cloud-enabled version of the ELK stack http://www.logsearch.io/. ELK is a tool for receiving, processing and outputting logs. ELK uses Elasticsearch as a back-end datastore, and kibana as a front-end reporting tool to create a powerful pipeline for storing, querying and analyzing your logs.
The Predix Logging service is based on several open-source software components:
- Elastic Search
- Indexes and stores the log data.
- ELK
- Receives and parses the log data.
- Kibana
- Provides a customizable, user-friendly way to visualize your log data.
Predix Logging Subscription Plans
Learn about your options for using the Predix Logging service.
Predix Logging subscription plans offer the following:
- Dedicated Single Node
- Preferred for higher memory with non-HA requirements. This could be used for production use cases if the application is designed for resiliency.
- Dedicated and Highly Available (HA)
- Best for production usage where fault tolerance in the Predix Logging service is required. These plans can also be used to accelerate read access to data structures as there are two read replica servers in the topology.
Details of each plan type follow.
Predix Logging Dedicated Single-Node Configuration
| Plan Feature | Feature Description |
|---|---|
| HA | Not included; the service will be interrupted upon any host reboot. |
| Persistence | ElasticSearch indices are writte to local storage. RDB Snapshots are enabled by default. The database will be saved when one of the following parameters has been reached:
|
| Backups | Hourly with 24-hour retention. |
| VM | Dedicated VM with single Predix Logging process; no noisy neighbor issues. |
| Recommendation | Recommended for advanced development or testing requirements, or in production where Highly Available configuration is not required. |
Predix Logging Dedicated Highly Available configuration
| Plan Feature | Feature Description |
|---|---|
| HA | Single Read-Write primary server replicating to two Read-Only replica servers; Redis/Sentinel provides fast, automatic primary server promotion and client failover when using Sentinel- aware clients. See the samples included in this document. |
| Persistence | ElasticSearch indices are sharded, replicated, and stored both locally and on a second node in the cluster. RDB Snapshots are enabled by default. The database will be saved when one of the following parameters has been reached:
|
| Backups | Hourly with 24-hour retention |
| VM | Dedicated VMs with single Predix Logging process each; clients can connect to the replica servers to accelerate read performance. |
| Recommendation | Good for production environments that require Highly Available configuration. |