Skip to content

Commit 2517879

Browse files
windergmalouf
andauthored
Update system requirements. (#1581)
Co-authored-by: Gary <982483+gmalouf@users.noreply.github.com>
1 parent e498ab7 commit 2517879

File tree

1 file changed

+5
-4
lines changed

1 file changed

+5
-4
lines changed

README.md

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -23,7 +23,7 @@ Run `make` to build Indexer, the binary is located at `cmd/algorand-indexer/algo
2323

2424
# Requirements
2525

26-
All recommendations here should be be used as a starting point. Further benchmarking should be done to verify performance is acceptible for any application using Indexer.
26+
All recommendations here should be used as a starting point. Further benchmarking should be done to verify performance is acceptible for any application using Indexer.
2727

2828
## Versions
2929

@@ -33,10 +33,11 @@ All recommendations here should be be used as a starting point. Further benchmar
3333

3434
For a simple deployment the following configuration works well:
3535
* Network: Indexer, Conduit, Algod and PostgreSQL should all be on the same network.
36-
* Indexer: 2 CPU and 8 GB of ram.
37-
* Conduit: 2 CPU and 8 GB of ram.
36+
* Indexer: 1 CPU and 1GB of ram. Scale up for systems with high query volume.
37+
* Conduit + Algod: 4 CPU and 8 GB of ram.
38+
* Storage: algod follower nodes, 40 GiB, 3000 IOPS minimum.
39+
* Deployments allocating less ram might work in conjunction with [GOMEMLIMIT](https://pkg.go.dev/runtime@master#hdr-Environment_Variables) for Algod (and even Conduit). This configuration is not tested, so use with caution and monitor closely.
3840
* Database: When hosted on AWS a `db.r5.xlarge` instance works well.
39-
* Storage: 20 GiB
4041

4142
A database with replication can be used to scale read volume. Configure a single Conduit writer with multiple Indexer readers.
4243

0 commit comments

Comments
 (0)