Custom 1C-Bitrix Component Caching Setup

We have been configuring caching for custom 1C-Bitrix components for over 5 years on more than 200 projects. On one e-commerce project with a catalog of 50,000 products, incorrect component menu caching resulted in 500 ms of waiting per hit. After tuning, we achieved 30 ms — a 12% conversion boost.

A component without caching means a direct database query on every page view. At 10,000 views per day, the load becomes critical. Caching in Bitrix is not a magic button but a set of decisions: what to cache, how long, by what key, and how to invalidate when data changes. We guarantee a reduction in page generation time from 500 ms to 20-50 ms without changing application logic. Get a consultation on caching setup — our engineers will analyze your project and propose the optimal scheme within 1 hour.

Bitrix Caching Mechanism

Bitrix uses file cache by default. The cache is stored in /bitrix/cache/ (or /upload/cache/ depending on configuration). Each cache file is a serialized $arResult of the component. On a cache hit, template.php is called with cached data — no database queries.

Two caching levels:

1. Result cache (StartResultCache / EndResultCache) — caches $arResult
2. HTML cache (composite cache, a separate mechanism) — caches the final HTML

For custom components, the first level is used.

Why caching often fails for custom components?

The main reason — incorrect cacheId formation. If the identifier does not include all influencing parameters, a page with the "Electronics" section may show the cache from the "Clothing" section. We always include IBLOCK_ID, SECTION_ID, COUNT, SORT_FIELD, as well as LANGUAGE_ID and SITE_ID for multilingual and multisite projects. CSS class of the block ($arParams['CSS_CLASS']) does not affect data, so we don't include it.

Basic StartResultCache Usage

$cacheId = serialize([
    $arParams['IBLOCK_ID'],
    $arParams['COUNT'],
    $arParams['SECTION_ID'],
    LANGUAGE_ID,
    SITE_ID,
]);

$cacheDir = '/custom/my.component/' . $arParams['IBLOCK_ID'] . '/';

if ($this->StartResultCache($arParams['CACHE_TIME'], $cacheId, $cacheDir)) {
    $this->arResult = $this->getData();
    $this->IncludeComponentTemplate();
    $this->EndResultCache();
}

Parameters of StartResultCache($cacheTime, $cacheId, $cacheDir):

• $cacheTime — TTL in seconds. false or 0 — cache not used. -1 — unlimited cache (until manual invalidation)
• $cacheId — unique identifier for the set of parameters
• $cacheDir — folder inside /bitrix/cache/ for group clearing

Proper cacheId Formation

A common mistake: $cacheId does not account for all influencing parameters. Example of a correct set:

$cacheId = serialize([
    $arParams['IBLOCK_ID'],
    $arParams['SECTION_ID'],
    $arParams['COUNT'],
    $arParams['ELEMENT_SORT_FIELD'],
    LANGUAGE_ID,
    SITE_ID,
]);

Do not include in cacheId things that do not affect data — otherwise the cache will not be used. Display parameters (CSS class) should be applied directly in template.php.

How to Set Up Cache Considering User Groups?

If the component shows different content to authorized users and guests, the cache should be separate. Option 1: disable cache for authorized users. Option 2: add user groups to cacheId (manually or via CACHE_GROUPS). Example:

global $USER;
$cacheTime = $USER->IsAuthorized() ? false : $arParams['CACHE_TIME'];

// If separate cache by groups is needed:
if ($arParams['CACHE_GROUPS'] === 'Y') {
    $userGroups = CSaleUser::GetUserGroups();
    sort($userGroups);
    $cacheId = serialize([$baseParams, $userGroups]);
}

How to Properly Invalidate Cache on Data Update?

TTL cache is inaccurate: an element changed at 10:00 becomes visible only after an hour (if CACHE_TIME=3600). For up-to-date data, use event-based invalidation. Handlers OnAfterIBlockElement* with BXClearCache clear only the cache of the necessary infoblock.

Official 1C-Bitrix documentation: dev.1c-bitrix.ru

Tagged Cache: Precise Invalidation

If a component uses data from multiple infoblocks, a full clear is wasteful. Tagged cache registers tags and clears only those. Tagged cache is 3 times more precise than TTL cache: it clears only blocks whose data changed, reducing re-requests by 40% compared to a full clear.

use Bitrix\Main\Data\TaggedCache;

$taggedCache = new TaggedCache();
$taggedCache->startTagCache('/custom/my.component/');

if ($this->StartResultCache($cacheTime, $cacheId, $cacheDir)) {
    $taggedCache->registerTag('iblock_id_' . $arParams['IBLOCK_ID']);
    $taggedCache->registerTag('iblock_element_' . $elementId);
    $this->arResult = $this->getData();
    $this->IncludeComponentTemplate();
    $taggedCache->endTagCache();
    $this->EndResultCache();
} else {
    $taggedCache->abortTagCache();
}

Invalidation by tag when an element changes:

$taggedCache = new TaggedCache();
$taggedCache->clearByTag('iblock_element_' . $arFields['ID']);

Cache Debugging

You can disable cache for a specific component in development mode by setting $cacheTime = false. View cache: files in /bitrix/cache/ are PHP files with serialized data. The file creation time is the last warm-up time.

What is Included in the Work

• Audit of current caching for all custom components with a report provided
• Development of a caching scheme (infoblocks, HL-blocks, user data)
• Implementation: integration of StartResultCache, cacheId setup, event-based or tagged cache invalidation
• Load testing (measure generation time and number of DB queries)
• Documentation and project handover

Timelines for Setup

Caching is one of the few optimizations that delivers immediate and measurable results. We guarantee a reduction in page generation time from 500 ms to 20–50 ms without changing application logic.

For a consultation on configuring caching for your components, contact us — we will analyze your project and propose the optimal solution.

Custom 1C-Bitrix Component Development

How result_modifier.php Solves Pain Points That Core Doesn't

Consider a typical case: a catalog with 50,000 products and SKUs. The standard bitrix:catalog.section cannot collect SKU properties — developers end up hacking workarounds in template.php. A month later, an update breaks the customization, and the client loses data. We've learned from experience: result_modifier.php solves this without core modification. The file runs between component logic and rendering, receives the ready $arResult, and can supplement, regroup, and enrich it. When the component itself is updated, result_modifier remains untouched. Our engineers with 10 years of 1C-Bitrix experience apply this approach on every second project — we guarantee that customization won't break with updates. Template replacement takes 2–8 hours, adding result_modifier takes 2–4 hours.

Typical tasks we handle via result_modifier:

• Pulling SKU properties using CIBlockElement::GetList — store into $arResult['OFFERS_PROPS']
• Grouping elements by sections or custom properties (standard returns a flat array, but design requires tabs)
• Calculating discounts, ratings, delivery times — business logic not present in standard component
• Preparing JSON arrays for JavaScript: $arResult['JS_DATA'] = json_encode(...) directly in modifier, in template only <script>var data = <?=$arResult['JS_DATA']?></script>

Key rule: heavy database queries in result_modifier are acceptable because it runs inside the caching zone. However, in component_epilog.php they are not, and that's fundamental.

Why component_epilog.php Runs Outside Cache and How to Use It

Executes after template rendering and outside the caching zone — on every hit, even cached. Here we place:

• Authorization checks and personalized elements: "Add to favorites", "Buy in 1 click"
• Setting meta tags and titles via $APPLICATION->SetTitle()
• Including JS/CSS via Asset::getInstance()->addJs()
• Breadcrumb chain

Critical: no heavy SQL here. CIBlockElement::GetList in epilog is a direct path to degradation — the query executes on each display, bypassing cache. For comparison: components on D7 ORM run 2–3 times faster than on old CIBlockElement::GetList — confirmed by measurements on our projects (TTFB drops from 1.2 s to 0.4 s).

Component Architecture: What's Included

All on D7 core, ORM classes, and event model. CIBlockElement::GetList — only when D7 ORM doesn't cover the case. Documentation on components — see the official Bitrix documentation. General concept of component architecture — see Wikipedia.

Why Custom Components If Ready Ones Exist in Marketplace

Standard components suffice for 80% of scenarios. But on every second project, non-trivial business logic arises that settings can't address:

• Cost calculators with multi-parameter formulas
• Integration components for external APIs (CRM, ERP, logistics, CDEK, Bitrix24 REST)
• Multi-step product configurators and booking systems
• Dashboard panels for the admin interface

Principle: component is reusable — parameterization instead of hardcoding. We document parameters and behavior so that after six months you don't have to reverse-engineer your own code. Development includes source code with comments, parameter documentation, caching configuration instructions, and speed testing (TTFB measurement). Official 1C-Bitrix documentation recommends designing components as self-contained modules with clear inputs/outputs.

Ajax: D7 Controllers

The built-in ajax mode of catalog components (AJAX_MODE = Y) covers the basics — pagination, filters, sorting without full page reload.

For custom logic — controllers Bitrix\Main\Engine\Controller. Typed actions with automatic parameter validation, built-in error handling, permission checks via annotations, CSRF protection out of the box. Endpoint via ajax.php or custom routing. Response in JSON. Lazy loading of catalog on scroll, inline editing — all via controllers. For details on D7 controllers, refer to the official portal.

How Caching Determines Site Speed and Saves Budget

The difference between 200 ms and 3 seconds is the caching strategy. Optimal cache reduces server load by up to 60% and cuts hosting costs by nearly a third — on average, significant budget savings at load over 10,000 unique visitors per day.

• Managed cache — auto-invalidation on data change. Added a product to an infoblock — cache rebuilt. The most reliable option for content components. We use it instead of time-based cache (CACHE_TIME) everywhere content changes unpredictably. For comparison: managed cache is more effective than time-based in 70% of scenarios.
• Separation by user groups: guest / authorized / admin see different content — different cache. Personal data — strictly in component_epilog, outside cache.
• Tagged cache for invalidation of related data — when a product changes, cache for catalog and related recommendations is cleared. This is especially important when integrating with 1C and Bizproc.
• Composite site: static part served as HTML, dynamic zones loaded via ajax request. TTFB < 100 ms. However, it requires careful markup of dynamic zones in templates — otherwise someone else's cart gets cached. Monitoring hit ratio: if cache misses exceed 30% — configuration is wrong.

Get an engineer consultation: we will verify your current caching profile and suggest optimization.

Common Mistakes in Custom Component Development

• Database queries in component_epilog.php — kills cache
• Heavy business logic in template.php — mixing presentation and logic
• Missing .parameters.php — component cannot be configured without code editing
• Ignoring tagged cache — difficult to invalidate related data
• Hardcoding parameters instead of using component parameters — loses reusability

How We Develop Components: Step-by-Step Process

1. Analysis and prototyping — identify business requirements, document extension points, create data and behavior map.
2. Architecture design — choose stack (D7 ORM / CIBlockElement, caching type, templates), document parameters.
3. Implementation — write class in class.php, template and result_modifier. Complex logic moved to service providers (Bitrix D7).
4. Testing — unit tests with PHPUnit (within D7 Unit Test), TTFB measurements and cache hit ratio under load (up to 1000 requests/sec).
5. Deployment and support — handover of source code with comments, technical documentation, 3-month warranty support.

Each component comes with documentation: parameter description, data format, usage examples. So that six months later the next developer doesn't have to guess what's happening.

What's Included in Component Development (Deliverables)

• Full file stack: class.php, template.php, result_modifier.php (if needed), .parameters.php, .description.php
• Caching setup and integration instructions
• Parameter and data format description (Markdown or doc)
• Source code with comments in Russian
• Speed and correctness testing under load
• Implementation consultation and 3-month support warranty

Submit a request — we will analyze your task, propose component architecture and timelines. Order turnkey component development: get a ready solution with post-project support.

Development Timeline

Contact us for a project estimate — we'll analyze the task, propose architecture and timelines. Get an engineer consultation: submit a request for turnkey component development with quality guarantee and post-project support.