PGSync : Real-time Integration Tool For PostgreSQL And Elasticsearch

Rate this post

PGSync is a middleware for syncing up information from Postgres to Elasticsearch easily. It permits you to maintain Postgres as your origin of truth and uncover organized denormalized records in Elasticsearch. Changes to settled substances are engendered to Elasticsearch. PGSync’s high-level query developer at that point creates upgraded SQL questions on the fly, depending on your outline. Elasticsearch training will get to know more about the real-time Integration tool for PostgreSQL and Elasticsearch.

PGSync : Real-time Integration Tool For PostgreSQL and Elasticsearch


PGSync’s advisory model permits you to rapidly move and change huge volumes of information rapidly while keeping up social integrity. Depict your report construction or pattern in JSON, and PGSync will persistently catch changes in your information and burden it into Elasticsearch without composing any code. PGSync changes your relational information into an organized report format. It permits you to exploit the expressive force and adaptability of Elasticsearch straightforwardly from Postgres.

You don’t need to compose complex inquiries and transformation pipelines. PGSync is lightweight, adaptable, and quick. Elasticsearch is more fit as an optional denormalized search engine to go with a more customary standardized datastore. Besides, you shouldn’t store your essential information in Elasticsearch. Apparatuses like Logstash and Kafka can help with this assignment; however, they require a touch of designing and development. Extract Transform Load and Change information capture instruments can be unpredictable and need costly designing exertion.

Different advantages of PGSync include:

  • Scale on-request
  • Real-time analytics
  • Effectively join various nested tables
  • Solid primary datastore/source of truth

You have information in a Postgres database at a significant level, and you need to reflect it in Elasticsearch. This implies each change to your information (Insert, Update, Delete and Truncate articulations) should be repeated to Elasticsearch. From the outset, this appears to be simple, and afterward, it’s most certainly not. Add some code to duplicate the information to Elasticsearch subsequent to refreshing the data set (or supposed double writes). Composing SQL inquiries traversing various tables and including numerous connections are challenging to compose. Distinguishing changes inside a nested document can likewise be very hard. Assuming your information never showed signs of change, you could take a preview on schedule and load it into Elasticsearch as a one-off activity.

Top Three Database Management Trends [2021]

PGSync is suitable for you if:

  • Elasticsearch is your read-only search layer, and Postgres is your read/write source of truth.
  • Your information is continually evolving.
  • You have to denormalize relational data into a NoSQL data source.
  • You have existing information in a relational data set. For example, Postgres, and you require an auxiliary NoSQL database like Elasticsearch for text-based inquiries or autocomplete inquiries to reflect the current information without having your application execute dual write.
  • You need to maintain your current information immaculate while exploiting the pursuit capacities of Elasticsearch by uncovering a perspective on your information without trading off the security of your relational data. Or on the other hand, you need to uncover a perspective on your relational information for search purposes.

How does it function?

PGSync uses the coherent decoding highlight of Postgres (presented in PostgreSQL 9.4) to catch a constant stream of progress events.

PGSync is written in Python (supporting version 3.6 onwards), and the stack is made out of Elasticsearch, Redis, Postgres, and SQlAlchemy. PGSync uses the coherent decoding highlight of Postgres (presented in PostgreSQL 9.4) to catch a constant stream of progress events. This element should be empowered in your Postgres configuration document by setting in the PostgreSQL.conf record:

> wal_level = logical

You can choose any pivot table to be the root of your document.PGSync’s inquiry developer assembles advanced inquiries progressively against your schema. PGSync works in an event-driven model by making triggers for tables in your database to deal with notice events. This is the lone time PGSync will at any point roll out any improvements to your database.

If you alter the structure of the PGSync’s schema config, you would require to rebuild your Elasticsearch indexes. There are plans to back up zero-downtime migrations to execute this method.


There are various methods of installing and executing PGSync

  • Running in Docker
  • Manual configuration

Running in Docker

Running in Docker is the simplest method to get up and running.

To startup all services with Docker. Run:

$ docker-compose up

Display the content in Elasticsearch

$ curl -X GET HTTP://[elasticsearch host]:9201/reservations/_search?pretty=true

Manual configuration

  • Setup

Ensure the database user is a superuser

Enable logical decoding. You will also have to set up at least two parameters at PostgreSQL.conf

wal_level = logical

max_replication_slots = 1

  • Installation

$ pip install pgsync

Make a schema.json for your report representation

Bootstrap the database (just one time) bootstrap –config schema.json

Execute the program with pgsync –config schema.json or as a daemon pgsync –config schema.json -d

Features of PGSync

Key features of PGSync are:

  • Effectively denormalize relational information.
  • Operates with any PostgreSQL database (version 9.6 or later).
  • Irrelevant effect on database execution.
  • Conditionally steady yield in Elasticsearch. This implies that the writes show up just when they are focused on the database, refresh, insert and erase activities show up in a similar request as they were submitted (instead of inevitable consistency).
  • Issue open-minded doesn’t lose information, regardless of whether cycles crash or any network interference happens, and so forth. The interaction can be recuperated from the last designated spot.
  • Returns the information straightforwardly as Postgres JSON from the database for speed.
  • Supports composite primary and foreign keys.
  • Supports a subjective profundity of settled elements, i.e., Tables having a long chain of related conditions.
  • Supports Postgres JSON information fields. This implies: we can extricate JSON fields in a database table as a different field in the subsequent report.
  • Adjustable report structure.

Benefits of PGSync

  • PGSync is easy to use the box answer for Change information capture.
  • It handles information erasures.
  • PGSync needs little improvement exertion. You characterize a composition config depicting your information.
  • PGSync produces progress inquiries coordinating with your scheme straightforwardly.
  • PGSync permits you to handily modify your records if there should arise an occurrence of a schema change.
  • You can uncover just the information you need in Elasticsearch.
  • Supports different Postgres constructions for multi-tenant applications.
  • When utilize PGSync diminishes the intricacy of most application stacks.
  • Postgres is your read/write source of truth, while Elasticsearch is your read-only search layer.
  • You have information that is continually evolving.
  • You have information in a current relational database, for example, Postgres, and you require an optional NoSQL database like Elasticsearch for text-based or auto-complete inquiries.
  • You need to keep away from the improvement overhead and intricacy commanded by different apparatuses.

Market Alternatives

  • Kafka: This is an Apache project that is an open-source stream-preparing software platform.
  • Kinesis: Amazon handled administration like Kafka.
  • Logstash: This is a result of Elastic which gathers information from different sources; at that point, parses, changes, and ships to different objections.
  • ZomboDB: Postgres augmentation that gives full-text search through the utilization of Elasticsearch records.


You have information in a Postgres database at an undeniable level, and you need to uncover it in Elasticsearch. This implies each change to your information should be repeated to Elasticsearch. At first, this appears to be simple, and afterward, it’s not. You were adding some code to duplicate the information to Elasticsearch subsequent to refreshing the database or executing the dual writes at your application level. Composing SQL inquiries crossing various tables and including different connections can be nontrivial.

Author Bio

VarshaDutta Dusa

I am VarshaDutta Dusa, Working as a Senior Digital Marketing professional & Content writer in HKR Trainings and having good experience in handling technical content writing and aspires to learn new things to grow professionally. I am expertise in delivering content on the market demanding technologies like ServiceNow Training, Oracle Service Bus Course, SQL Server DBA Course, Elasticsearch Course, Jmeter Course, Kibana, ServiceNow HR Service Management, etc.

Other References:

Introduction to pgsync: real-time integration tool for postgresql and elasticsearch (

Share and Spread the Love
About Shout Me Crunch 1035 Articles
Shout Me Crunch provides the latest technology news and views. We also provide the tech guide by video review or Step by step tutorial.

Be the first to comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.