Apigee cost vs self-hosted Apache APISIX: where the three-year math tends to land
A buyer facing an Apigee renewal, an Apigee Edge for Private Cloud end-of-life date or a rising call volume needs to know one thing first: self-hosted Apache APISIX removes per-call and subscription charges, but it adds infrastructure, build, developer portal and staffing costs that you must estimate yourself. APISIX is licensed under Apache-2.0, so the software itself carries no license fee. Apigee's charges, by contrast, are set by Google's commercial terms, and the only reliable figures for your organization are the ones in your own contract, including any environment, subscription and call-based line items. Pull those from the contract and the invoices before comparing anything.
Model the three years by component. On the Apigee side, the recurring lines are the Google-managed or hybrid subscription and the per-call and environment charges, which scale with traffic and the number of environments you run. On the APISIX side, the one-time items are the migration of proxies and policies, custom plugin work for Apigee policies that have no built-in equivalent, and the developer portal you add yourself. The recurring items are the compute and storage for the gateway and its etcd configuration store, observability, support if you buy it, and the platform engineers who patch, upgrade and on-call the stack. APISIX ships with more than 100 built-in plugins, which reduces custom code, but the count of plugins does not tell you how many of your policies they cover.
The result depends on two inputs you control: call volume and team size. High traffic makes per-call charges heavy, but it also raises infrastructure spend, so the gap narrows only if your hosting is efficient. A small team that already runs Kubernetes and etcd may absorb APISIX operations cheaply, while a team with no platform staff will pay for that skill in hires or consultants. Ask any implementation partner for a written scope that separates one-time migration, recurring infrastructure, portal, support and staffing, and check that each line names the assumptions behind it. Confirm the current APISIX release line with the project (3.19.0 was released on September 28, 2026) before you pick a version to standardize on.
Which cost components sit on each side, and which are one-time or recurring
This table sets the Apigee cost components against the self-hosted APISIX components, with each row marked as one-time or recurring and showing who is paid. Use it to build your own three-year model, then ask each vendor or partner to price the rows that apply to your environment. The main takeaway is that APISIX removes the gateway license line but moves the runtime, portal, upgrade and operations work onto your budget.
| Cost component | Who is paid | One-time or recurring |
|---|---|---|
| Apigee subscription, per-call and environment charges | Google, under your Apigee contract (confirm current terms with Google) | Recurring, typically tied to contract term and usage |
| Hybrid or Private Cloud runtime costs | Your cloud or data center provider for Kubernetes or servers, plus Google for hybrid management plane access | Recurring, scales with runtime nodes and environments |
| APISIX software license | No license fee; Apache License 2.0 (Apache-2.0) | None for the Apache-2.0 software itself |
| Infrastructure for APISIX data and control planes | Your cloud or data center provider, plus etcd storage and load balancers | Recurring, driven by traffic volume and high availability design |
| Implementation and migration of proxies and policies | Your internal team or an implementation partner, under a written scope | One-time, driven by proxy count and custom policy rewrites |
| Self-added developer portal | Your team or a partner for build and hosting; portal tooling may carry its own license | One-time build, then recurring hosting and content upkeep |
| Support subscription (optional) | The APISIX community is free; commercial support comes from a vendor such as API7.ai, under its own terms | Recurring if you buy it; confirm scope and response terms |
| Upgrades, security patching and testing | Your internal staff or a retained partner for ongoing support and testing | Recurring, tied to APISIX release cadence and your change process |
| Internal skills and on-call operations | Your payroll or contractor fees for platform engineers and API owners | Recurring, the largest line item for most self-run gateways |
What moves the comparison most for your organization
Apigee and self-hosted Apache APISIX price different things, so the same workload can land far apart depending on a few drivers. The cards below describe how each driver pushes cost on each side and what to measure in your own environment before you ask for a quote.
API call volume
Apigee's commercial charges are tied to how much traffic passes through its managed or hybrid runtime, so growth in calls moves the subscription side directly. APISIX has no per-call license, so volume shows up instead as compute, memory, network egress and logging storage on your own nodes. Measure peak requests per second, average and peak daily calls, and payload sizes for your top ten APIs, then ask Google for the call-based tier that matches those figures.
Number of environments
Each Apigee environment and environment group is a separately scoped commercial unit, so adding development, test, staging and production copies multiplies the license footprint. With APISIX, extra environments mostly cost you more etcd clusters, gateway nodes and pipeline runs, which your team controls and can size down for non-production. List every environment, its purpose and its traffic, and confirm with the vendor how each one is counted.
Number of APIs and proxies
Apigee bills and administers proxies in a way that grows with the number of API products, developers and apps, so a large catalog raises both the subscription and the admin overhead. In APISIX the cost of more routes is mostly migration and maintenance effort, since routes and plugins are configuration that reloads without restarts. Count proxies, shared flows, API products and consumer apps, and note how many change each quarter.
Availability and multi-region needs
Apigee hybrid keeps the runtime in your Kubernetes but depends on Google for the management plane, so high availability design has to account for both sides. A self-hosted APISIX cluster adds your own redundancy work, including multiple gateway nodes, etcd quorum across zones and tested failover between regions. Define your uptime target, the regions you must serve and the outage scenarios you must survive, then price the extra nodes and staffing those scenarios require.
Depth of Apigee policies needing custom plugins
Standard Apigee policies map cleanly to built-in APISIX plugins, but policies that rely on custom Java callouts, complex JavaScript or Apigee-specific extensions need rewriting. Each custom plugin is one-time engineering in your own language and then recurring maintenance on every APISIX upgrade. Inventory every policy type in your proxies, flag the ones with no direct plugin equivalent, and estimate the rewrite work per flag with your implementation partner.
Skills already on the team
A team that already runs NGINX, etcd and Kubernetes can operate APISIX with modest new training, while a team fluent in Apigee's UI and policy model will spend time learning configuration as code. Skills gaps shift cost into hiring, training or an outside retainer, and that cost recurs for as long as the platform runs. Survey who will own on-call, upgrades and security patches, and decide whether you need internal staff, a support contract or a retainer for the three-year model.
How to build the three-year comparison with your own numbers
This sequence turns your Apigee contract, your own traffic measurements and written quotes into a three-year total cost for both options, so the comparison reflects your estimates rather than generic figures. Each step produces a record that the next step depends on.
Collect the current Apigee invoices and contract terms
Finance and the account owner pull the last twelve months of Apigee invoices, the order form or subscription agreement, and any amendments. Record the billing unit, the environment and subscription charges, any included call allowance, overage rates, add-ons such as analytics or portal licensing, and the renewal date. Ask Google or your reseller to confirm in writing any term that is unclear, such as how calls are counted and which products are billed. The step is done when every line of the current spend maps to a named contract clause.
Inventory proxies, environments and monthly call volume
The platform team exports the list of API proxies, shared flows, products, developer apps and environments, using the Apigee management API or apigeecli where the team already has access. Pull monthly call counts from the Apigee analytics or billing reports for each environment and each API product, and separate production from non-production traffic. Note which proxies use custom policies, JavaScript, Java callouts or external service dependencies, since these affect migration effort. The step is done when you have a table of proxies, environments and monthly calls that matches the billing report within a range you can explain.
Load test APISIX to size nodes on your own traffic
Stand up an APISIX test environment in your own infrastructure using the deployment mode you plan to run, traditional, decoupled or standalone, and replay representative traffic from the inventory at peak and at expected growth levels. Measure throughput, latency percentiles, CPU and memory per node, and etcd behavior if you use etcd. Count the nodes, etcd members and load balancer capacity needed to hold your peak with headroom. The step is done when you have measured requests per second per node for your plugin set and a node count that you can defend in writing.
Price the implementation scope in writing
Define the migration scope with the implementation partner or internal team: which proxies move, which Apigee policies need custom APISIX plugins, how the portal and analytics are rebuilt, and how environments are cut over. Request a written quote that lists deliverables, assumptions, client responsibilities and exclusions, with each deliverable tied to an inventory item. Ask vendors to state whether a fixed-scope or time-and-materials model applies and what changes would trigger a new quote. The step is done when every proxy in the inventory is either inside a priced deliverable or explicitly excluded.
Add portal, analytics, support and staffing costs
Estimate the components Apigee bundles that APISIX does not provide out of the box: a developer portal you select or build, an analytics stack such as Prometheus, Grafana or ClickHouse, identity integration, backup, and support contracts for the chosen components. Estimate the staffing hours for platform operations, on-call, upgrades and security patching, and record who pays for each hour, whether internal payroll, a retainer or a support contract. Separate one-time costs, such as build and migration, from recurring costs, such as infrastructure, licenses and operations. The step is done when each component has a named owner, a cost source and a one-time or recurring label.
Build the three-year totals for both options
Create a spreadsheet with one row per cost component and one column per year for the Apigee path and the APISIX path. Use contract figures for Apigee, measured node counts and quoted implementation fees for APISIX, and your own staffing rates for operations. Include a parallel-run period in which both platforms carry traffic, and include any Apigee exit or notice obligations stated in the contract. The step is done when both totals sum to the same time horizon and every cell links back to a source document, measurement or quote.
Run a traffic growth sensitivity check
Repeat the three-year totals at several traffic growth assumptions, using the growth rates from your own call history and business plan rather than a generic figure. Test the Apigee overage and tier thresholds against each scenario, and test APISIX node counts against the same peaks from the load test. Identify the growth level at which the cost curves cross, if they do, and note which inputs move that point most. The step is done when you have a short table showing both totals at each growth scenario and a named list of the inputs that a reviewer should challenge.
Why migration cost is one-time while APISIX run cost recurs
The move from Apigee to self-hosted Apache APISIX splits into two kinds of cost. Migration is a bounded project: exporting proxy bundles, rebuilding policies as APISIX routes, upstreams and plugins, writing custom plugins for Apigee policies that have no direct match, and moving the developer portal. Once those deliverables are accepted, that spend ends. Infrastructure, upgrades, security patching and the people who operate the gateway recur every year for as long as the platform runs. Ask any proposal to show those two lines separately, so the one-time work does not hide the permanent run rate.
The infrastructure line depends on the APISIX deployment mode. In traditional mode, the data plane and control plane run together and etcd sits alongside them, so you size and monitor one cluster plus the etcd members it needs. In decoupled mode, data plane nodes and control plane nodes scale separately while sharing one etcd cluster, which adds components to run but can cut the cost of scaling traffic. In standalone mode there is no etcd, and configuration comes from a file or the Admin API, which lowers the platform footprint but moves change control into your own pipelines. Confirm with the APISIX deployment modes documentation which mode fits your traffic and change model before pricing hardware.
Plan for a dual-run period. While traffic moves environment by environment, Apigee subscription or environment charges continue alongside the new APISIX infrastructure, and your staff run two gateways at once. Those overlaps are temporary but real, and they are often the largest single surprise in a first-year budget. The end-of-life date for Apigee Edge for Private Cloud 4.53.01, June 30, 2027, as stated in Apigee documentation, is a timing driver only: it sets when a cutover or an upgrade decision must be made, and it does not change the cost of the platform itself. Ask your vendor for the contract end date and the notice terms for your Apigee deployment so the overlap window is planned, not guessed.
How the comparison changes by situation
If your APIs run on Google-managed Apigee, the bill is driven mostly by usage and by the environments and tiers you subscribed to. Your contract, not a list price, sets the per-call and subscription terms, so start by pulling the order form and the most recent invoices. Compare the line items against what APISIX would need to run the same traffic.
- Call volume: Confirm the billed call count per month and how it has grown over the last year, since per-call charges scale with traffic.
- Environments and add-ons: List every non-production environment, analytics or monetization add-on, and support tier, because these often recur independently of traffic.
- Self-hosted run cost: Estimate the compute, load balancers, etcd and logging for APISIX at your peak and average volumes, using your own cloud pricing.
- Build and migration effort: Ask for a written scope covering proxy export, policy mapping and cutover, since this is one-time cost that a short comparison often omits.
- Operating staff: Identify who would patch, upgrade and on-call the gateway, and whether that time already exists in your platform team.
Before comparing totals, test whether your call volume is stable or spiky, because a spiky profile changes the break-even point.
In Apigee hybrid, Google hosts the management plane while your team runs the runtime plane on your own Kubernetes clusters. Your infrastructure and operations costs are already visible, so the question is how much of the Apigee subscription would stop once the runtime moves. The runtime infrastructure you already pay for may be reusable for APISIX, which changes the comparison.
- Cluster reuse: Check whether the Kubernetes clusters and ingress capacity that run the hybrid runtime can host APISIX without new hardware or licenses.
- Management plane charges: Ask which part of the subscription covers the Google-hosted management plane and whether it stays payable if you move only some APIs.
- Operator time: Measure how much platform engineering time the hybrid runtime already consumes, including upgrades and certificate rotation.
- Data residency value: If the hybrid split was chosen for residency, price the cost of losing that control in your review rather than assuming it is free.
- Partial migration: Ask whether you can move a subset of APIs first while the hybrid subscription continues, since that shifts cost into a parallel-run period.
Confirm with Google which contract terms apply to a partial exit and how long the hybrid subscription must run after a notice of change.
Apigee Edge for Private Cloud has a fixed lifecycle. Apigee documentation states that version 4.53.01, first released on August 26, 2025, reaches end of life on June 30, 2027, after which Apigee stops providing bug fixes, security updates, hotfixes or patches for that line. Continuous support requires running a version within its 18-month full support window. The cost question therefore includes a forced decision, not just a renewal.
- Upgrade path: Compare the effort of upgrading within the Edge line against a move to APISIX, since an upgrade extends the existing dependency rather than removing it.
- Support continuity: Confirm with Apigee which support window applies to your installed version and what unsupported operation would cost you in risk.
- Infrastructure refresh: Identify the servers, Cassandra, Postgres or other stores that Edge for Private Cloud depends on, and whether they need replacement regardless of the gateway choice.
- Parallel running: Budget for running old and new gateways side by side during cutover, which is a one-time cost that a steady-state comparison hides.
- Retirement cost: Include decommissioning, data retention for logs and audit records, and archival of proxy bundles.
Ask for the support and end-of-life terms in writing, and confirm the dates against Apigee documentation before you set a budget.
Questions buyers ask before comparing Apigee and APISIX costs
How much effort does moving from Apigee to APISIX take?
Effort depends mainly on how many proxies, shared flows and custom policies you run, and on how many of those policies have a direct APISIX plugin equivalent. Start by inventorying each Apigee proxy, its policies, its target endpoints and its consumers, then sort them into three groups: those that map to built-in APISIX plugins, those that need a custom plugin, and those that depend on Apigee features with no open-source counterpart. Ask any implementation partner to price the work by that inventory and by the number of environments you cut over one at a time. Your own environment, such as identity provider integration and CI pipelines, will shift the estimate as much as the proxy count.
Can an internal team run Apache APISIX without outside help?
Many platform teams can run APISIX in-house if they already operate Kubernetes or Linux services, etcd and NGINX-based infrastructure, and they have someone who can write and test plugins. APISIX supports traditional, decoupled and standalone deployment modes, so you can choose the one your team can operate. The skills gap usually sits in observability, upgrade testing and incident response, not in routing configuration. Assess your team against those skills before deciding, and confirm the exact operational steps with the APISIX documentation and your own security team.
What should a quote for an APISIX migration include?
A useful quote lists each deliverable and the environments it covers. Ask for these items as separate lines:
- An inventory and mapping of Apigee proxies, shared flows and policies to APISIX routes, upstreams and plugins
- Any custom plugin development, with the languages and test approach agreed
- Identity provider integration, for example through the openid-connect plugin
- Deployment, hardening, backup and upgrade procedures in your infrastructure
- Cutover plans, rollback steps and knowledge transfer to your team
- Documentation and any ongoing support or retainer, shown as a separate item
DeployOpen works on fixed-scope engagements with scope, deliverables and timelines agreed in writing before work starts, and it does not publish a price list.
Why should a three-year cost model cover three years?
A three-year window captures the one-time migration cost and several years of recurring infrastructure, staffing and support, which is where most of the difference between platforms appears. Apigee pricing is tied to usage and subscription terms, so a short model can hide growth in call volume and environments. A model that is too short understates the migration cost, and one that is too long relies on assumptions about traffic and staff you cannot verify. Confirm the period with your finance team, and confirm Apigee's current charges with Google and your contract, since this page does not supply those figures.
Do we need a third-party support subscription for APISIX?
APISIX is licensed under Apache-2.0, so you can run it without a license fee. Whether you need a paid support subscription depends on your internal skills, your uptime requirements and how much you want an outside party to answer for production incidents. The supplied research does not list the commercial support options for APISIX, so ask vendors such as API7.ai, which offers an enterprise product, for their current terms in writing. Compare any subscription with the cost of an on-call internal team, and remember that a retainer for upgrades and testing is a separate line item.
Which costs stay the same when we leave Apigee?
Several costs move to your own budget rather than disappearing. You still pay for the infrastructure that runs the gateway, the etcd cluster in traditional or decoupled mode, logging and monitoring, and the people who maintain them. A developer portal is not included with APISIX, so you must add one, and that work and its hosting are a real cost. Map each Apigee charge to the component that replaces it, then list the new components you must build or buy, and check the result against your security and compliance obligations.
Which APISIX version should we plan around for production?
Plan around a version your team can test and upgrade on a schedule. The APISIX downloads page lists version 3.19.0, released on September 28, 2026, and it does not mark any release as long-term support. Because no LTS designation is published, confirm your upgrade cadence, the support window you accept and the testing you will fund before you fix a version. Check the current release notes on the official downloads page at the time you plan.
When self-hosting APISIX is cheaper and when it is not
Self-hosting APISIX tends to pay off when call volume is high, steady and predictable, and when several environments and teams all need the gateway. Apigee's commercial charges scale with traffic and with the number of environments and subscriptions, so a large estate with production, staging, development and partner sandboxes pays for each of them. APISIX's Apache-2.0 license carries no per-call fee. Your costs become the servers or Kubernetes capacity, the etcd cluster, and the engineers who run them. Those costs grow in steps rather than with each request, which is why the balance tips toward self-hosting as traffic and environment count rise. The reader should price the infrastructure at peak and at idle, not at an average, because a gateway must be sized for its busiest hour.
For a small estate, managed Apigee can cost less than it first appears. A team with a handful of APIs, modest traffic and no platform engineers may find that the analytics, developer portal and monetization features Apigee includes would otherwise need to be built, and that building them costs more than the subscription. APISIX ships with more than 100 built-in plugins covering authentication, traffic control, observability and AI gateway functions, but it does not include a developer portal or business analytics, so those are work you add yourself. Ask Google and any partner for a quote covering your actual call volume and environment count, then compare it with a written scope for APISIX, including portal and reporting work.
Plan your own upgrade cadence before you commit. The APISIX downloads page does not mark any release as long-term support, so no vendor-designated support window exists for you to rely on. Apigee's managed service removes much of this patching work, while a self-hosted gateway moves it to your team. Decide who tracks new APISIX releases, how quickly security fixes go into staging and production, and how often the etcd cluster and the Ingress Controller are upgraded. Budget for that recurring work in the three-year model, and confirm with your own security and change teams how frequently they will accept gateway upgrades.
How DeployOpen scopes a fixed-price APISIX implementation
DeployOpen is a Chainscore Labs company that plans, implements, hardens, documents and supports self-hosted open-source platforms inside your own infrastructure. For a move from Apigee to Apache APISIX, we start with a written scope that separates the cost components you actually control from the ones you do not.
How we scope
- We inventory the Apigee environments, proxies, products and portal features that are in use today.
- We agree in writing which APISIX deployment mode fits your hosting, network and etcd requirements.
- We list the Apigee policies that need custom plugins, so build effort is visible before work starts.
- Engagements are fixed-scope, so deliverables and timelines are agreed in writing before any work begins.
How we build
- We stand up APISIX in your cloud accounts, data centers or air-gapped network, with configuration kept in version control.
- We rebuild routes, upstreams and authentication through the openid-connect plugin against your existing identity provider.
- We test traffic behavior, rate limits and failover in a staging environment before any production cutover.
- We integrate logging and metrics so traffic reporting continues while you decide what business reporting still needs.
How we hand over
- You receive the documented architecture, configuration repository, runbooks and a list of every rebuilt Apigee policy.
- We hand over an environment-by-environment cutover record, including rollback steps for each stage.
- We document the costs that recur after handover, such as infrastructure, etcd operation and any licensed portal you add.
How we support
- Ongoing support, upgrades and continuous testing are available as retainers after handover.
- We can prepare systems and evidence for compliance assessments, while independent auditors issue any formal reports.
- Ask us for a quote built from the written scope, with each cost component itemized so you can compare it with your Apigee renewal.
