Ignition Edge historian limitations

I understand that Edge has the limitations of 1 week of data storage for trending.

But does that mean that an Edge client is also limited to viewing only 1 week of data even when it is connected to a full gateway which is logging that data?

Basically trying to figure out if I can have an edge site that is connected to a full gateway, but can view more than 7 days of trends if it’s connected to the central gateway (e.g.pulls the data back from the central gateway to view)?

If there is a link between Edge and the main Gateway, then to view more than 1 week of data the client at the Edge NUC should be launching the full Gateway client.

Also note that it is 1 week or 10M rows, whichever may come first.

Also to note if MQQT module is added, then that removes the 10M row limitation.

That. If you want to use Edge to ‘buffer’ history data locally, but a central server for main record keeping, then that’s explicitly a supported architecture; just launch your clients against the ‘main’ server, and they’ll natively query history. If you also want to be able to run a project from your Edge gateway locally (in event of a network failure, for instance) then with Edge panel you can configure ‘Local Client Fallback’; any Vision client running on the same machine as the Edge gateway will automatically switch to running a client off the local gateway in the event it loses connection to its “primary” gateway.


Thanks Paul.

My main concern is around the bandwidth as these remote sites could have as low as 300kbps connections. I figured running a client against the main gateway (probably in the cloud) at these speeds would be less than ideal (note that it’s for viewing trend data only, the actual control would come from another vendors panel located on the site).

Unless those speeds are ok for things like trend viewing?

In terms of data logging there’d be up to 250 tags being trended (on change or 15 sec intervals)

Once the client’s downloaded, there shouldn’t be a large volume of network traffic; just whatever data needs to be sent to keep the screen updated (realtime tags, and trend data). The gateway network is fairly efficient about that transfer, although there’s enough variables that the only way to know for sure would probably be to actually test things :slight_smile:

Ok sounds good.

So if I understand correctly… I can have a remote site that under normal conditions runs a client that is against the main gateway. Whilst this is happening the edge gateway is connected to my local devices and buffering and sending data through to my main gateway.
Then in the event the remote site loses connection to the main gateway, it will automatically close the client from the main gateway and launch a client from the edge gateway?

Yep, exactly. Local client fallback is an automatic function of any running Vision client; the ‘master’ gateway (in your scenario) doesn’t know anything about it or require any configuration.

Sorry to bring this up again, but is this “1 week” limitation still the case as of version 8.1.x? Or is it really the 10Million records the real limitation? Hypothetically, If I have only 100 total records stored and the connection has been lost for 1 month…I get nothing when connection is re-established? I may be wrong, but limiting data storage based on timestamp is irrelevant compared to number of records.

Pretty sure you get nothing. Edge has always been targeted at applications with no need for local history, except for short-term communications outages. If you haven’t deployed a central gateway to store history for your Edge installations, I think you are doing it wrong.

1 Like

Might not be a concern for you, but just in case, be aware that there is an additional limitation with the Edge Historian. Unlike the full Tag Historian, the Edge Historian does NOT track and validate scan classes. So be careful with how you set up your tag history on the tags and pull/aggregate the data. You want to avoid settings that allow for varying time gaps between data points or come up with a custom method for validating your data (heartbeat, etc). If you don’t you can get some really inaccurate results for periods where the Edge gateway goes down or other glitches.

Edit: Just realized this topic is over a year old, whoops!

By all means, I completely understand and we did have a central gateway. We had a case where the remote site was on Cellular connection and snow/ice covered the antenna, and a snow storm prevented anyone from going on-site for more than 1 month. In that case I feel like Edge could at least just keep 10 million points regardless of time frame due to acts of God events.

Pretty sure force majeure (think I spelt that right) is everyone’s get-out clause, and IA would be no exception

Why not have the local device store data to something such as SD-Card?

I have all of our “Edge of Network” device’s writing certain data to a local SD-Card for just such a problem. I have it setup to store 90 days at which point it will create a new “FILE”.

I have several times (Due to weather) logged in and using FTP pulled this file back to my computer and then manually updated my SQL database in the Ignition Server.

It’s isn’t a question of technology. It’s a question of IA’s license. Edge doesn’t do databases or long-term storage. If you want that, buy a full Ignition license. Note that a full license can be just the platform for $1k, which allows database connections, full gateway scripting, and the most popular OPC drivers. (That’s about at the price point for Edge w/ gateway scripting. I would say the full license is by far the better deal. )

1 Like

I brought the topic back because we want to utilize Edge…and 10 million records is just fine. It’s the 1-week part that seems too restrictive in cases where someone can’t visit the remote site due to natural disasters.

Yes, in terms of pricing, the full version is starting to make more sense. We’ve been able to get our remote projects to not even utilize gateway scripting to help fit in the lowest cost of Edge (just using Sync). We used to use the full Ignition license and enjoyed having a few setup screens created in Vision to be used only in the Designer when setting up a site. With Edge, as soon as you activate the license, Vision won’t run in trial mode or in the Designer. Ignition still allows this. So, we might begin migrating back to create PC images with Ignition instead of Edge.

Full Ignition also allows you to use 3rd-party modules. I put Simulation Aids and Life Cycle everywhere–they are just too useful. But they are not on the Edge whitelist.

You’re good. My post was in response to the original topic creator for “edge limitations” and I missed that it was an old post.

Full is nice but I understand and agree with the limitation being called out here. And Full gateway pricing doesn’t compare when you also need Emerson ROC and MQTT Transmission module (along with its other dependencies). Start adding those modules and the costs begin to add up really fast and when you have hundreds of edge locations it’s not viable. The basic short term tag historian on edge is great but I don’t think we should stop requesting enhancements. It’s a new product that needs to grow just like all platforms.

I also just learned today that the MQTT transmission module publishes based on the same tag value change that the edge historian uses. We were hoping to store at a rate similar to the poll rate, but publish at a much slower rate. Unfortunately that isn’t a native feature yet either.