STATUS
Is Heroku Connect going away?
Heroku has not published an end date for Heroku Connect, and its public statement is narrower than the rumors. In a February 6, 2026 post, Heroku announced a sustaining engineering model: no new features, and no Enterprise contracts for new customers. Existing subscriptions and support contracts stay honored. The post does not mention Heroku Connect by name.
For most teams that makes this a planning question. If your app is leaving Heroku, or your Heroku Connect renewal is due, the sync layer has to move too. Heroku Connect only writes to Heroku Postgres, so it is often the last thing holding a company on Heroku.
For the background on what Heroku Connect does and where teams outgrow it, read our explainer on why teams switch from Heroku Connect. This page covers the buying decision: pricing, Salesforce API usage, hosting, the other alternatives and the cutover.
PRICING
How pricing compares with Heroku Connect row tiers
Buyers we spoke with described Heroku Connect plans sized by row count. The common complaint was the gap between the tier and the rows that matter: plans sized for hundreds of thousands of rows that synced a small fraction, and tiers in the tens of millions of rows where one to two million records changed in a month.
Stacksync bills on two things: records in sync and active syncs. The pricing page defines the first one: "A record becomes a synced record when it is present at the beginning of the month or when it syncs for the first time." Updates to a synced record are unlimited.
- Starter: $1,000 a month billed annually, 50K records in sync, 1 active sync.
- Pro: $3,000 a month billed annually, 1M records in sync, 3 active syncs.
- Enterprise: custom volume and unlimited active syncs, quoted.
Two worked examples
An oversized small plan. A team pays for a Heroku Connect plan sized for a few hundred thousand rows but syncs about 50,000. On Stacksync, those 50,000 records fit the Starter allowance. Compare $12,000 a year against your renewal quote, and size the plan on the records you sync today plus a year of growth.
A large tier with a small active set. A team pays for a tier in the tens of millions of rows while one to two million records change each month. Stacksync prices on records in sync, as the pricing page defines them, with unlimited updates. The lever is scope: sync the objects and records your app reads, such as open opportunities and active accounts, and keep closed history out of the sync. Above 1M records, the per-record rates on /pricing apply, and the largest volumes are quoted as Enterprise.
Plan features differ too. Sync log retention, security frameworks and support grow by plan; the comparison table on /pricing lists them. For a line-by-line look at Heroku's own costs, see our Heroku pricing breakdown.
SALESFORCE API USAGE
Will a replacement use your Salesforce API limits?
Yes. On API quota, Heroku Connect has the edge. Heroku's Dev Center documents that the API calls Heroku Connect makes to Salesforce don't count against your org's daily API request limit. Any third-party replacement, Stacksync included, reads and writes through the Salesforce API under your org's quota.
Plan for it before cutover. Enterprise Edition orgs get 100,000 requests a day plus 1,000 per user license. The first full backfill uses the most calls, so schedule it and watch it during the proof of concept. After that, usage tracks how many records change. Stacksync includes smart API rate limit handling on every plan. Our guide to Salesforce API limits walks through checking your current usage and estimating what a sync adds.
LATENCY
Polling vs real-time sync
Heroku Connect polls in both directions. From Salesforce to Postgres, it polls every 10 minutes by default, configurable from 2 to 60 minutes (Heroku Dev Center: mapping options), with accelerated polling through the Salesforce Streaming API or Change Data Capture (Heroku Dev Center: reading data). From Postgres to Salesforce, it reads its trigger log every 2 minutes, or as often as every 10 seconds when a pg_notify trigger signals a change (Heroku Dev Center: writing data).
Stacksync syncs changes in both directions in real time on Salesforce objects where it can use Apex triggers. Trigger mode needs an integration user with the Author Apex and Customize Application (or Modify All Data) permissions. Objects that can't take triggers, or a user without those permissions, fall back to incremental polling on a slower cycle. Confirm the mode for each object early and include the polled ones in the proof of concept.
WHERE YOUR DATA LIVES
Salesforce sync on the Postgres you choose
Heroku Connect needs Heroku Postgres. Stacksync connects to PostgreSQL wherever it runs:
- Amazon RDS for PostgreSQL inside your own VPC
- Render Postgres
- Supabase (see Salesforce and Supabase)
- A self-managed PostgreSQL server
- A database Stacksync manages for you (see database hosting)
Scope the project first. Stacksync replaces the Salesforce sync, not the Heroku runtime. Your dynos move on their own track, and copying existing tables out of Heroku Postgres is a standard Postgres dump and restore. The Salesforce and PostgreSQL integration page covers the connector itself.
FROM THE FIELD
What teams replacing Heroku Connect told us
These come from our calls with teams evaluating a switch. Names and identifying details are removed.
- Billing on the wrong unit. One buyer paid for a row tier in the tens of millions while one to two million records changed each month. Stacksync prices on records in sync, so the useful question became which records need to be in sync at all.
- Several connections, one database. One team ran several Heroku Connect connections against one Salesforce org and wanted them consolidated into one self-hosted Postgres.
- API quota as a hard requirement. One buyer made "no impact on our Salesforce API limit" a condition of the switch, because Heroku Connect never touched that quota. That requirement is why the backfill plan belongs in the proof of concept.
- Agencies moving whole books. An agency with many Heroku Connect clients asked two questions first: what the total cost is once Salesforce API usage is included, and how clients locked into Heroku contracts can get out.
CUTOVER
How a switch from Heroku Connect usually runs
Most cutovers follow five steps. The Heroku Connect migration guide covers each one in detail, from mapping to leaving Heroku Postgres.
- Inventory the mappings. Export the objects and fields Heroku Connect maps, and list the app queries, triggers and reports that read those tables.
- Set up the target. Create the Postgres database the sync will write to. Heroku Connect keeps running against Heroku Postgres in the meantime.
- Connect and backfill. Connect Salesforce and Postgres in Stacksync, map the same objects and fields, and run the first backfill while you watch Salesforce API usage.
- Run both in parallel. Compare row counts and spot-check records across both databases, including the slow objects and any write-back to Salesforce.
- Cut over. Point the app at the new database, turn off Heroku Connect writes, and cancel the remaining contract or ask whether it qualifies for a buyout.
