# Refreshing a datasource

Refreshing your datasource ensures that the data Tellius uses for analysis is current and accurate. You can:

* **Refresh Immediately**: Update the data on the spot.
* **Schedule a One-Time Refresh**: Set a specific future date and time to run the refresh once.
* **Configure a Periodic Refresh**: Establish a recurring schedule (daily, weekly, monthly, etc.) to keep data continuously up-to-date.

By choosing the right refresh strategy, you maintain data freshness while optimizing performance and reducing manual work.

### Refreshing a data source

1. Go to **Data → Connect**, click on the three-dot kebab menu of the required datasource and select **Refresh**. The following window will be displayed:

<figure><img src="https://977923713-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJHwf1QFuv1BRPzfSnL2Z%2Fuploads%2FRiMt9x1LLAocPog36rnh%2Fimage.png?alt=media&#x26;token=1beba720-c778-4702-a94a-457bd5ba6866" alt="" width="563"><figcaption><p>Refreshing a data source</p></figcaption></figure>

2. In the left pane, select **"All Datasets"** to refresh all datasets associated with the selected datasource. Else, select the individual datasets to refresh rather than the entire datasource.
3. Choose the refresh method from the three available tabs: Immediate, Periodic, One-time.
4. All the refreshes can be tracked in the **Notifications** tab.

#### Immediate Refresh

<figure><img src="https://977923713-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJHwf1QFuv1BRPzfSnL2Z%2Fuploads%2FLpyLu17KeuVvwo3k8El8%2Fimage.png?alt=media&#x26;token=f4c67229-32af-4837-a3e6-3da4d361dba9" alt="" width="563"><figcaption><p>Immediate refresh</p></figcaption></figure>

* Ideal if you have just updated the underlying data and need to see the changes reflected instantly for ad-hoc analyses or quick data validations.
* Click **Run** to start the refresh process immediately.
* Tellius fetches fresh data right away, ensuring you see the latest information in your insights and dashboards.

#### One-Time Refresh

<figure><img src="https://977923713-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJHwf1QFuv1BRPzfSnL2Z%2Fuploads%2FML0faCUXOBmlIMLd7ViG%2Fimage.png?alt=media&#x26;token=a26619e1-7292-4cf3-9ce8-73c781c67a39" alt="" width="563"><figcaption><p>One-time refresh</p></figcaption></figure>

* Useful if you want to ensure data is refreshed overnight or just before a big presentation, without setting up an ongoing schedule.
* Provide a **Name** for this one-time refresh task.
* Set the **Start Date, Start Time and Time Zone** to specify exactly when the refresh should occur.
* Click **Save** to schedule the refresh.
* Tellius will automatically refresh the dataset(s) at the specified time—just once. This can be tracked under **Notifications**.

#### Periodic Refresh (Recurring Schedule)

<figure><img src="https://977923713-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FJHwf1QFuv1BRPzfSnL2Z%2Fuploads%2FE69JUfw75ScDpIQKSV5h%2Fimage.png?alt=media&#x26;token=db317883-938c-4fce-b65b-ea9a1e01204e" alt="" width="563"><figcaption><p>Periodic refresh</p></figcaption></figure>

1. Ideal if you want daily or weekly refreshes to maintain continuously fresh data for ongoing analytics. Ensures data is always up-to-date without manual intervention.
2. Enable **TurnOnPeriodic Refresh** by toggling the switch.
3. Choose a **Frequency** (e.g., daily, weekly, monthly) from the dropdown.
4. Set the **Start Date**, **Start Time**, and **Time Zone** for the first occurrence. The dataset will refresh at this frequency starting from the given date and time.
5. Provide a **Name** for the refresh schedule.
6. Click **Save** to confirm.
7. Tellius sets up a recurring schedule. The datasource refreshes automatically at the chosen intervals, keeping your analysis current without manual triggers.

### Best Practices

* If you’re unsure of your latest data’s availability, run an immediate refresh first. This confirms that everything works before scheduling.
* If refreshes take longer than expected, consider adjusting the frequency or refreshing fewer datasets at once to optimize performance.